Skip to main content

Full text of "Yaksas"

See other formats




t^'ision of Ethnology 
J.S. National Museum 



(With 50 Plates) 


Keeper of Indian, Persian, and Muhammadan An 
Museum of Fine Arts, Boston 

(Publication 3059) 

MAY 19, 1931 


be A 





(With 50 Plates) 



Keeper of Indian, Persian, and Muliammadan Art 

Museum of Fine Arts, Boston 




(Publication 3059) 

MAY 19, 1931 




(With 50 Plates) 


Page 2, note i, add: Arbman, E., Riidra, Upsala, 1922. M. Peri (B. £. F. 
E. O., XVII, ii, p. 46) has remarked in connection with ball offerings (bhuta- 
yajfia) of all kinds that they " present a family likeness which leads us to 
suppose that they are all, whether Buddhist or Brahmanical, derived from ancient 

Page 4, note 3, add : Walhouse, J. M., On the belief in bhutas, devil and ghost 
worship in Western India, Journ. Anthrop. Soc, 1876, pp. 411, 412; and Wick- 
remasinghe, Don M. de Z., Cat. Sinhalese Mss. in the British Museum, 1900, 
pp. 44-54. 

Page 5, substitute for the first and second paragraphs : 

The word Yaksa occurs several times in the Rg Veda, Athari'a Veda, Brah- 
manas, and Upanisads; in the earlier texts it has generally been thought to mean 
" something wonderful or terrible," not clearly definable. In the most recent 
discussion, Hillebrandt ^ finds as the earliest meaning " magician, uncouth being, 
unseen spiritual enemy, etc.," then simply a " supernatural being of exalted 
character," and finally " Yaksa " in the ordinary sense. The etymology of 
" Yaksa " is very uncertain ; Professor A. B. Keith writes to me that he regards 
it as obviously connected with the root yaj, to worship; Hillebrandt suggests a 
connection with Vedic yaks in pra-yaks, to honor. There may be a connection 
with the mysterious fever called yaksma in the Atharva Veda. Or the word, 
as well as the source of the concept, may be non-Aryan. 

In any case the ideas of- the wonderful, mysterious, supernatural, unknown, of 
magical power, invisibility, and spirit-hood are all more or less involved in the 
early references ; but these ideas are hardly to be distinguished from those con- 
nected with the Yaksa concept when later on the cult of Yaksas comes clearly 
into view, and it is often, especially at that time, difficult to distinguish between 
Deva, Devata, and Yaksa, especially in the Buddhist literature, where all alike 
are regarded as rebirths of human beings, and subject in due course to further 
human incarnation. In one place or another every Indian deity without excep- 
tion, is spoken of as a Yaksa, and in all these cases the sense is honorific. 

In the earliest texts a dual attitude is recognizable, one of fear and dislike, 
the other of respect. The first seems to me to reflect merely an Aryan dislike 
and distrust of aboriginal deities. For this attitude we have the texts RV., IV, 

* My Yaksas, Washington, 1928, is now out of print. The Addenda here made 
(amongst which those to be substituted for two paragraphs on page 5 of the 
original text are by far the most important) will be incorporated in a new edition ; 
in the meantime they are made available to those who possess the first. 

^Hillebrandt, A., Vedisch Yaksa, in Festgabe Richard Garbe, Erlangen, 1927. 


3, 13 "Do not (O Agni) consort with the Yaksa (? famiHar spirit) of any 
smooth swindler, intriguing neighbour, etc." ; RV., V, 70, 4 " Let us not, O ye gods 
of great power, encounter a Yaksa"; RV., VII, 56, 16 yaksadrso "espying the 
Yaksa" (regarded as an invisible enemy to what is being undertaken) ; RV., 
VII, 6, 15 where yaksa in the sense "invisible" seems to be contrasted with 
citra in the sense of " visible " and Kausika Sutra, 93, 3, where Yaksas are 
classed with other adbhutani as creatures of ill-omen. Charpentier, in J. R. A. S., 
1930, pp. 325-345, argues that the word ndicasakha in RV., Ill, 53, 14, means 
" worshippers of the banyan tree " and that the cult was hateful to Aryans be- 
cause of human sacrifices performed in connection with it (for a contrary view 
see ib., p. 894). 

Before discussing the second attitude, that of high respect, characteristically 
exhibited in the Atharva Veda and Upanisads, reference must be made to 
RV., I, 24, 7, "In the Unsupported (sky) King Varuna, he of purified intelli- 
gence, sets up the top of the tree. Downward are they (the branches), above 
their base. May the rays reside in us " ; and RV., X, 82, 5 " Prior to the sky, 
prior to this earth, prior to the living gods, what is that Germ which the 
waters held first and in which all the gods existed? The waters held that 
same Germ in which all the gods existed ; on the navel of the Unborn stood 
that in which all beings stood." This prototype of the later (Mbh. Ill, 272, 
44 and XII, 207, 13) conception of the reclining Narayana, resting upon the 
waters, and giving birth to Brahma (demiurge) by way of a lotus, of which 
the stem rises from his (Narayana's) navel, is developed as follows, in the 
Atharva Veda, X, 7, 38, with reference to Varuna, Brahman or Prajapati 
as the supreme and ultimate source of life : "A great Yaksa in the midst of the 
universe, reclining in concentrated-energy (tapas) on the back of the waters, 
therein are set whatever gods there be, like the branches of a tree about a 

Significance is to be attached to this concept of the tree of life springing from 
a navel.^ For Yaksas are primarily vegetation spirits, guardians of the vegetative 
source of life (rasa = sap in trees =: soma = amrta), and thus closely connected 
with the waters.^ These ideas of the origin of life in the waters are set forth in 

^ See my Tree of Jesse, and Indian parallels or sources, Art Bulletin, XI, 2, 
1929. The creative significance of the navel appears also in Avestan mythology 
in connection with Apam Napat = Vedic Apam Napat, the " son of the waters," 
who is also Agni, the word napat meaning both " offspring of " and " navel of," 
cf. Zend ndfyo, " offspring," from ndfa, " navel." On the sexual significance of 
the navel (ndbhi) and vedi as both representing the womb of the Earth Goddess 
see Johannsen, Ueber die altindische Got tin Dhisand, pp. 51-55. A relation or 
identity of Apam Napat with Varuna is suggested by RV., II, 35, 8 and V, 2, 8 
where all creatures and plants are described as shooting out from Apam Napat 
and multiplying in progeny. Cf. below, p. 24, note. 

*Cf. RV., VII, 65, 2 and 88, 6, also Dlgha Nikaya, II, 204, where Varuna is 
called a Yaksa; AV., XI, 2, 24 "Thine, O Pasupati, is the Yaksa within the 
waters, for thine increase flow the waters of heaven"; Vessavana's (Kubera's) 
sea Dharani "whence arise the clouds, whence the rain falls" {Dlgha Nikaya, 
II, 201) ; the powers exercised by the tutelary Yaksa Sata (p. 3, infra) ; and the 
connection of the Yaksas and Yaksis with makaras and other riverine monsters 
as " vehicles." 


the later " decorative " art of the water-cosmology by the constantly recurring 
formula of a lotus rhizome bearing leaves and flowers {lata-kamma, mala-kamma, 
Vin. 117, 152, cf. lata-yatthi, kiisuma-yatthi, Mhv. XI, 10-13), often supporting 
or framing birds and animals (cf. sakuna-yatthi, Mhv., loc. cit.), and typically 
springing from the mouth or navel of a Yaksa, or the more obvious water 
symbols, the brimming vessel (punna-ghata) or open jaws of a makara or a fish- 
tailed elephant. A fuller treatment of this subject. will be found below; here we 
need only remark a connecting link between the Yaksa, who is Brahman and the 
later Yaksas who are Lords of Life. That Yaksas in the accepted sense are 
familiar to the Atharva Veda is more definitely established in VIII, 10, 28 where 
Kubera and his son are already called punyajana (also ib. XI, 10, 24), "good 
folk" (itarajana, "other folk" in the Kashmir text), and what they are said to 
" milk " from Viraj as their subsistence is the power of concealment. In A.V., 
XI, 6, 10 Yaksas are invoked with all the other gods mentioned in the same 

To speak then of Brahman or Prajapati as a great Yaksa is effectively to say 
of him, great divinity, great power. Nor is this the only Vedic passage (though 
it may be the earliest) in which he is thus called Yaksa. In the Gopatha 
Brahmana, I, i and TCiitt. Brahmana, 3, 12, 3, i Brahman speaking says "by 
concentrated energy (tapas) I became the primal Yaksa"; in the Brhadaranyaka 
Upanisad, 5, 4 we have " He who knows that great Yaksa as the primal-born, 
that is, that Brahma is the Real, he conquers these worlds " ; in Kena Upanisad, 3, 
(15), and Jainiiniya Upanisad Brahmana, IV, 20, where Brahman shows him- 
self to the gods, who know him not, they ask "What Yaksa is this?," and 
come at last to know through " Uma " that it is Brahman. In all these cases, 
no doubt the implication "wonderful being" is involved; but we come* much 
nearer to understanding the native plexus of ideas if we retain the original 
" Yaksa," than we do by a translation expressly designed to avoid a sup- 
posedly mistaken identification of the Vedic and the later Yaksa concept. As 
a matter of fact, as we have already shown, the Yaksa concept in its " later " 
sense was certainly known to the Atharva Veda, and therefore and so much the 
more was it certainly well-known in the time of the Upanisads. In reality the 
Aupanisadic " Brahman-Yaksa " represents a concept that goes back to the Kg 
Veda, where it is originally applicable to Varuna. An intermediate stage is repre- 
sented in a passage of the Atharva Veda (X, 3, 43) where the indwelling spirit 
or self of man is called atmanvat Yaksa — " The lotus flower of nine gates, veiled 
by the three qualities (gunas), what self-like Yaksa dwells therein, that (only) 
the Brahman-knowers know." 

For Yakkha as individual soul see also Pali Text Soc, Pali Dictionary, s. v. 
Yakkha, 7. As atmanvat and as arakkha devatd, cf. Persian Fravashi. 

In the Grhya Sutras (Gobhila, 3, 4, 28; Asvalayana, 3, 4, i ; Sankhayana, 4, 93) 
Yaksas are invoked as hhutani; Kubera with Isana, for the husband in the 
marriage ritual (Paraskara, I, 8, 2; Sankhayana, I, li, 7). 

Bharhut inscriptions make all four of the Maharajas or Regents, and not only 
Kubera, into Yakkhas. 

Page 5, note 2, read : The word Bhuta might be rendered " those who (were 
originally men, and) have (now) become (spirits)." Add: Cf. Arbman, E., 
Rudra, Upsala, 1922, pp. 165 ff., where the use of the word is discussed at 


Page 6, line 21, add: Kubera's wife is called Bhadra, "Lucky" in Malta- 
bharata, I, 199, 6, and Rddhi, " Success," ih., XIII, 146, 4, f . ; he is also said to be 
united to Laksmi {ib.. Ill, 168, 13, etc.) but she is here to be regarded mainly in a 
general sense as the goddess of Fortune, who associates herself with all great 
kings. In the Buddhist legends about Kubera as Pancika, his wife (Nanda, 
Abhirati, Hariti) has a much more definite character. In the later art the form of 
Hariti is closely assimilated to that of §ri-Laksmi. I have mislaid a reference 
to Kubera as goddess of a garden. 

Page 6, paragraph 3, continue : Jataka No. 281 tells of the great mango tree of 
Vessavana Maharaja, which grows on the Golden Mountain in the Himalayas, 
and is guarded by Kumbhanda-rakkhasas ; and in Jataka No. 489, where the 
rope trick is described, the tree magically produced is called " Vessavana's 
mango." The Indian rope trick is described in detail in early Celtic literature, 
where it is attributed to Manannan Mac Lir, god of the sea (Standish Hayes 
O'Grady, Silva Gadclica, London, 1892, p. 321) : Cf. Celtic "good folk" and 
" hidden people " with Indian punyajana and giihyaka designating Yaksas. 

Page 6, note 2, add : Statues of Kubera arc to be set up in the crypt of a 
treasury, Kautilya, Arthasdsfra, II, 4, see Meyer, p. 75. 

Page 6, note 3, add: The jewel-bearing lotus rhizome of the Bharhut coping 
relief has its source in the mouth of a kneeling elephant, not in this case a fish- 
tailed elephant ; the idea expressed is very clearly conveyed by a passage in the, 
of course much later, Dasakiimaracnnta of Dandin, Ch. X, where a girl is com- 
pared to a "jeweled vine (ratana-manjartkd) from the wishing-tree in Paradise 
{nandana-kaJpa-in-ksa) plucked by the sky elephant (Atrdvafa) and tossed 
to earth." Nandana, of course, is equally the paradise of Indra and of Kubera 
(Hopkins, Epic mylhology, p. 84), though Airavata (a rain-cloud) is primarily 
Indra's elephant steed (perhaps once Varuna's). See pi. 77, figs. 1-3. 

Page 7, line 19, add: In the Vidhura Pandila Jataka Punnaka is Senapati 
(General) and nephew of Vessavana. 

Page 7, line 29, after addition, add: but the conception of Ganesa in his 
original capacity of causer of hindrances to Vv^orship (piijd-vighnakartrnam) 
occurs frequently. 

Page 7, end of third paragraph, add: Sir Charles Eliott informs me that in 
Japanese Buddhist temples where Ganesa is worshipped the offerings made 
include spirits (sake), although as a rule alcohol is strictly forbidden in such 
places, and this affords further evidence of Ganesa's Yaksa connections. 

In several places there are indications of a connection of the goddess Sri 
(Laksmi) with Yaksas, e. g., as wife of Dharma in the Epic, she is the mother 
of Kamadeva, and her hand bears the mark of a makara, his symbol; in some 
Chinese Buddhist texts (Peri, loc. ctt., p. 39) she is the daughter of Manibhadra. 
In Jataka No. 392 she is the daughter of Dhatarattha, who is a Yaksa at 
Bharhut. See my Early Indian iconography, II. Srl-Laksmt, in Eastern Art, 
Part 3, 1929. 

Page 7, note 2, add: According to Burgess and Indraji, Inscriptions from the 
cave temples of Western India, 1881, p. 87, there was originally at the end of the 
verandah of Cave XVII at Ajanta a painting of a royal figure, with the inscrip- 


tion Manihhadra. Yaksas play important parts in Dhanavala's Bhavisatta Kaha 
(ed. Jacobi, 1916), where Manibhadra is mentioned (p. 13*) ; and in Hari- 
bhadra's Sanatkumaracarita (ed. Jacobi, 1921). 

Page 7, last paragraph, add footnote : In the Rdmayana, Tadaka, originally a 
YaksinI, becomes a RaksasT, and this is a great fall. In the Petavatthu Attakatha, 
no the devatta bhava of Yakkhas is contrasted with the pctatta bhava of Petas. 

Page 8, note i, add : For the cult of Kubera in Java, see Krom, Archaeological 
description of Bdrdbudur, pp. 17, 18, and for Yaksas generally in Java, see 
Stutterheim's review of Yaksas, in Djawa, 9^ Jaargang, No. 4 en 5, 1929, 
pp. 283, 284. 

Page 9, for the whole of the last paragraph, substitute : 

The case of HaritI or Harltl is too well known to need a long discussion. The 
best and fullest account occurs in the Samyuktavastii, Ch. XXXI (see Peri, N., 
Harm, la Mcre-dc-Demons, B. £. F. E. O., XVII, iii, 1917). We get the 
following genealogy : 

Paficala Sata 

(Yaksa king in Gandhara) (Tutelary Yaksa of Rajagrha) 

s. Paficika — j d. Abhirati s. Satagiri 

(= Kubera) j (rriHariti) 

500 Yaksa children (includ- (succeeds Sata as guardian deity 
ing Priyahkara or Pihgala, the of Rajagrha) 

Before her birth as a YaksT, Abhirati had been a herdsman's wife in Rajagrha, 
Because she had been required to dance at a festival while pregnant, she con- 
ceives the desire to avenge herself ; and now, despite the protests of Satagiri, she 
constantly devours the children of Rajagrha. The people make offerings to 
appease her; these consist of food, perfumes, flowers, cleansing and decorating 
the town, and making music, but matters are not improved. Then the guardian 
of Rajagrha makes it generally known by sending dreams that the only help is 
to be had from the Buddha. The latter hides Abhirati's youngest child under his 
begging bowl ; she is distracted by the loss, and searches in vain. The Buddha 
points out to her the moral, converts her, restores the child, and promises that 
offerings of food shall be regularly made to her in the monasteries, of which 
she becomes the protectress. 

I Tsing mentions that Hariti's image used to be painted in Indian monasteries 
near the refectory door, though this does not seem to have been de rigeur. 
The Samyuktavastu mentions only paintings of Yaksas to be thus made. Hariti 
is said to have given her children to the Samgha ; but they had to receive food 
at other than the regular hours, and even unclean food. 

As a popular divinity the converted Hariti was extensively worshipped as a 
giver of children; we may say, that having had her complex cured by the great 
master of psychology, she reverts to the normal. Examples are cited by Peri, 
loc. cit., pp. 65, 66; and according to M. Foucher's observations (B. £. F. E. O., 


I, p. 342, and Sur la jronticrc Indo-Afghanc, pp. 194-197) her cult in this 
sense still survives amongst the Muhammadans of the North-West. 

The monastic offerings seem to have been made originally on an altar set 
before painted icons of HaritI and Atavaka (Alavaka) placed within the re- 
fectory; later upon an altar out of doors. It is particularly interesting to find, 
though only from a late Chinese source (Peri, loc. cit., pp. 55, 56) that this 
external altar was made of stone in the form of a lotus flower, expanded towards 
the sky, and with its smooth round center serving as a table ; because precisely 
such altars have been found in Ceylon (Colombo Museum Guide, PI. Ill; here 
PL VI, 2). 

It will be observed that in the case of Abhirati, an evil wish conceived when a 
human being is fulfilled in the person of a malevolent YakkhinI ; exactly parallel 
to this is the case of the cannibal YakkhinI of the Ayoghara Jataka, No. 510, 
where the barren co-wife of the king of Benares, jealous of her fertile sister, 
prays to be able to devour the latter's children, and when reborn as a Yakkhini 
is able to fulfil this desire. 

Page 9, note i, add: Haribhadra, 11, 8. 2 (Leumann, E., Die Avasyaka- 
Erzdhhingen, Leipzig, 1897) tells a story in which a Yaksa bestows on a human 
artist the ability to depict the whole of a figure, though only the smallest part 
of the body may have been actually seen {ex pede Hercidem!). 

Cf. also the story of the production of a simulacrum of the Buddha by Mara 
at the request of Upagupta, Divyavadana, LXXVII and Asokavaddna, VII (see 
my Origin of the Buddha image, p. 42, and Pryzluski, J., Agokavaddna, 1923, 
p. 361). 

Page 10, second paragraph, add: Another Assamukhi is represented amongst 
the rocks of the Inda-sala guha on the Mathura lintel, B 208 in the Lucknow 
Aluseum, here obviously only as a " part of the scenery." 

Cf. the old legends of the mare forms of Saranyu and Vac (Keith, Rel. and 
Phil, of the Veda, pp. 198, 199). 

Page 10, note 3: in Bana's Kadamharl, 240, 241, Kinnaras haunting the forest 
are described as horse-faced. Generally speaking, however, both in literature and 
art, Kinnaras are of the siren type. 

Page ii, line 4, add fresh paragraph: 

The Milindapahha (191) has a list of cults, mentioning followers (gaiias) of 
Manibhadda, Punnabhadda, Candima, Surya, Siri-devata, Kali-devata (v. 1, 
Kali), Siva, and Vasudeva (v. 1, Vasudeva), adding that "the secrets of each of 
these sects are handed on in the sect itself, and kept hidden from all others." 
The Sirhhalese commentary calls the followers of these divinities bhaktas. In 
the Niddesa {Culla-Niddesa, pp. 173, 174) list of theistic cults, Manibhadda- 
deva is mentioned with Punnabhadda-deva, Yakkhas generally, and Maharajas 
(Regents) ; all four of the latter are Yaksas at Bharhut. Yaksas are mentioned 
in lists of deities in the Maitri Upanisad, I, 4, and vii, 6 and 8. In the Kdutillya 
Arthasastra, Ch. 25, there are to be shrines (kosthaka) for various deities, 
including Vaisravana, within the city (but kosthaka in this sense is unparalleled 
and perhaps granaries for the storage of grain from temple villages is really 


Page 12, note 2, after Agni, add : as Siva. There is also a goat-faced form of 
Skanda, connected with the procreation of children (Hopkins, Epic Mythology, 
pp. 229, 230). For Skanda as a phallic deity, cf. Meyer, J. J., Sexual life in 
ancient India, pp. 560, 561. 

^ Page 13, Inie 17, add new paragraph : 

In the Jatakas, the Bodhisattva is very often born as a tree-spirit, who displays 
characteristic virtues ; but the tree-spirit in such cases is always called a Devata, 
the Buddhist tendency being to restrict the designation Yakkha to demons, al- 
though there are many places where Devata and Yakkha are synonymous, c. g., 
Jat. No. 347. Dcva and Yakkha are distinguished in Jat. IV, 107: but Sakka 
(the Bodhisattva) takes the form of a Yakkha and appearing to his former wife 
teaches her the law. In S. N. i, p. 54, yakkha =^ dcva put ta. 

Page 13, note i, add: In the Sutasoma Jdfaka (for an illustration sec my 
Mediaeval Sinhalese art, figs, 151, 152, here pi. 25, fig. 4), and corresponding 
passage of the Dhammapada Atthakatha (2, 14 f., = Book V, story i) the non- 
cannibal character of a tree-spirit is very strongly emphasized; for when a king 
mistakenly attempts to offer human sacrifices to the tree, the rukkha-dcvatd 
exerts himself to the utmost to avert the catastrophe. Similarly in Jataka No. 9. 
Cf. Sata's objection to Abhirati's behavior, supra, p. 5. 

It is also to be observed that the Bodhisattva, whose births are always favor- 
able, is frequently represented in the Jatakas as having been born as a tree-spirit, 
and then often as behaving with great magnanimity. 

Page 15, transfer paragraph following line 28 (" In the .... exhausted ") to 
page 14, after line 17, adding: We have in Jataka No. 473, and in the case of 
Trikotti-Boyu just cited, specific examples of Yaksas reborn as men. In fact, the 
idea of alternate human and spirit birth — the idea, in fact, of sainsdrd — seems 
to be inseparably bound up with the Yaksa theology. 

Insert new paragraphs in place of the one transferred: 

In Hsiian Tsang's translation of the Mahd-vibhdsa-sastra (cited by Peri, in 
B. E. F. E. O., XVII, iii, p. 32) the Yaksa Madhusugandha says, " In my former 
life I was always your friend. Now I am born amongst the divine company of 
the Four Maharajas. I live on the gate Jivaka and protect the people." 

In the Vidhurapandita Jataka, No. 545, the king's mother at the time of his 
last birth but one previous to the present had been his guardian angel (drakkha- 
devata) and now again aids him in the same capacity. 

The case of the Yaksa Sata, who in the HaritI legend as related in the 
Samyuktavastu, Ch. XXXI (Peri, loc. cit., pp. 3f.) protected the people of 
Rajagrha, and after his death was succeeded by his son Satagiri in the same 
benevolent capacity, is especially significant, because the functions involved in 
protection are given in unusual detail (they are similar to those which might be 
expected to be fulfilled by Indra himself, or which result from the virtue of a 
king). Sata protected Bimbisara and all his house, and it is by Sata's grace that 
rain fell in due season, plants flourished, lakes were full, and there was no famine 
in the land ; he protected, too, the ascetics, Brahmans, the poor, the orphans and 
the merchants who flocked to the prosperous land of Magadha. 


Page 15, note i, fourth line, after -mentary, add: also Jataka, VI, 411; and 
Uvasaga Dasao, VII, 187, where the deva is sa-khinkiniyaim, explained by 
Hemacandra as meaning " wearing a girdle set with small bells." 

Page 16, after the last line, fresh paragraph : 

We have cited above two particular instances of Yaksa dvdrapalas or guardians 
oi gates. And as we know, Yaksas are constantly represented in this capacity, 
on either side of the entrance to a shrine.' It may be assumed that practically 
every building had to be protected by a spirit guardian. But accident could not 
always be relied on, and it is evident that the necessity for providing such 
guardians underlies an old form of human sacrifice which has survived, at least 
in tradition, into modern times. That human beings have been sacrificed and 
laid in the foundations of buildings is well known. It was popularly believed that 
such sacrifices were made when the Hooghly Bridge was built ; and a reviewer 
of my Yaksas (Modern Review, Nov., 1928) remarks that in Bengali legends 
misers are said to entomb little boys alive, and " the boy is presumed to take the 
form of a Yaksa — known as a Yaksa in Bengal — and stand guard over the 
treasure." Cf. Crooke, W., Popular religion and folk-lore of northern India, 
pp. 23yf{., and Index; and Bates, Hindi Dictionary, s. v. jak. An instructive 
early example is found in Jataka IV, 246 (No. 481) : "A great gate (dvara) 
is possessed and guarded by great spirits (devatd). A Brahman .... must 
be killed, his flesh and blood must be offered as a bali, and his body laid beneath, 
and the gate raised upon it." 

Page 17, first paragraph, add: In Jataka No. 398 the Yakkha Angulimala, 
originally a tree-dwelling cannibal (it is interesting to observe that the shadow 
of the tree is the limit of his power), when converted and reformed, is "given a 
seat" (nisldipctva) at the city gate; this most likely implies the establishment of 
a shrine and statue, and perhaps a daily cult, but might only mean the planting of 
a tree and setting up of an altar. Cf. Jataka I, 169, where the devatd of a 
banyan tree at the gate of a village receives bloody offerings. 

Page 17, line 15, add: Quintus Curtius (VIII, 9), speaks of capital punish- 
ment inflicted for injury done to a sacred tree. 

Page 18, after line i, add: Patafijali, commenting on Panini II, 2, 3, 4, 
refers to the sounding of musical instruments at gatherings in the temples 
(prasdda) of Dhanapati (Kubera), (Bala-) Rama, and Kesava (cited, Bhan- 
darkar, Sir R, G., Vaisnavism . . . ., p. 13). It would seem evident that the 
Besnagar kalpa-vrksa (here pi. I), a banyan tree with pots and bags of money, 
and a lotus and conch exuding coins, at its roots, must have been the capital 
of the dhvaja-statnbha of a temple of Kubera. 

Page 18, line 2 from bottom, for pediments read: tympanums. 

Page 19, note 4, add : The Aupapdtika or Ovavdiya Sutra, may be dated about 
the second century B. C. (Barnett, Hindu gods and heroes, p. 91). 

Page 19, note 5 : for court read caurl. 

Page 20, note 3, add : See also my Picture showmen, in Indian Hist. Qtly. 
V, 1929. 


Page 22, last paragraph ("Another story .... Mahabrahma "), transfer to 
top of page 26. 

Page 23, Hne 22i, delete (uttaraslsakam) . 

Page 23, after line 37, add: The prevalence of Yaksa cults in Magadha 
is further indicated in Jataka 307, where the Bodhisattva being born as a 
palasa-rukkha-devatd, deity of a palasa-tree, it is remarked " at that time all 
the inhabitants of Benares were devoted to the worship (inangalikd) of such 
devatas, and constantly engaged in religious offerings and the like (balikaran- 
adisu)." A similar statement is made in the Dummcdha Jataka, No. 50, and here 
the Bodhisattva seeing a crowd of worshippers with bloody offerings at a Banyan 
tree, " praying to the dcvata who had been reborn in that tree to grant them sons 
and daughters, honor and wealth, according to their hearts desire," himself draws 
near and behaves as a worshipper (piljctvd), offering perfumes and flowers, 
lustration (abhiscka), circumambulates {padakkhhmm katvd) and so honors the 
devata, though actually only with a view to the ultimate substitution of a higher 
faith for that of the popular tree-cult. 

Again, in Jataka No. 347, with reference to the past, we find " at that 
time men made hali offerings to the Devatas. The Bodhisattva forbade the 
killing of animals for these offerings; then the Yakkhas, losing their halikamma, 
were enraged." As in so many other cases, so here Dcvata = Yakkha. 

Page 24, after first paragraph : The problem of the relation of a tree-spirit 
to a tree is of some interest. With very rare exceptions it is a spirit in the 
tree, not the tree itself that wills and acts, as explained in Milindapanha, 
IV, 3, 20: "'the aspen tree conversed with Bharadvaja.' But that last is said, 
O king, by a common form of speech. For though a tree, being unconscious, 
cannot talk, yet the word ' tree ' is used as a designation of the dryad who 
dwells therein, and in that sense * the tree talks ' is a well-known expression." 
In general, the life of the tree-spirit is independent of that of the tree. Nothing 
is commoner, when trees are to be cut down, than a request to the tree-spirit 
to move elsewhere, the request being accompanied by offerings. In the 
Mayamataya it is said that a branch of the tree should be broken off, laid on 
the ground, dragged away and placed under another tree; evidently to provide 
the spirit with a necessary connecting link. Tree spirits, however, in the 
Jatakas and elsewhere are often spoken of as leaving their trees temporarily 
and appearing elsewhere on various occasions, so that they cannot be regarded 
as inseparably attached to their homes. On the other hand, in Jataka 398, the 
domain of a tree-spirit seems to be limited to the area covered by the shadow of 
the tree. Gifts are placed on trees, or on the altar at the foot of a tree. In 
consecrating trees or groves, the adhibasa (invocation) ceremony is performed 
on stone platforms below them (Agni Purana LXX). 

To find a Naga, instead of a Yaksa, as a rukkha- devata, as in the Mahdvanija 
Jdtaka, No. 493, is altogether exceptional. 

Page 24, after second paragraph, add : Offerings to Yaksas are generally called 
bali, and although this term is used equally of offerings made to all the gods, and 
also to the Manes, it is to be understood, when specially called an offering to 


All Beings or All Souls (sarvatmabhuti) , as in Manu, III, 91, as especially 
intended for the Yaksas. In Mahavamsa, XXXVI, 82, f., the cannibal Yakkha 
Ratakkhi is converted, and bali offerings are allotted to him at the entrances to 
villages. Bhntayajna w^ould appear to be the same thing as halt. 

Page 24, note 2, add : Barua's identification of Yaksa shrines with hero mounds 
(I. H. Q., II, 1926, p. 22) is similarly mistaken; so too de la Vallee Poussin, 
in L'Indc mix temps dcs Maury as, pp. 149, 150. 

On Yaksa caityas see also Chanda, Mem. A. S. I., 30, pp. 3 ff. Cf. also 
Atharva Veda, IX, i, the word cay. 

Page 25, line 10, after Pisacas, read : in Jataka No. 353 the Yakkha of a banyan 
tree receives bloody offerings. 

Page 25, line 13, add footnote to Ajanta: * See Goloubew, Ars Asiatica, 
X, pp. 19-20. 

Page 25, after line 30, add : In Haribhadra's Avasyaka-hka, II, 8, 2 ff. the 
Jakkha Surappia (Surapriya) has a temple at Ayodhya; the image must be 
repainted annually, and if this is neglected, he inflicts a pestilence on the people. 

Page 26, story of Buddhi and Siddhi, add note : Cf. Manimekhalai, Bk. XXII 
(transl. by S. K. Aiyangar, p. 169). 

Page 2y, line 2, after " Mayana," read: (=Kamadeva). 

Page 27, after line 3, fresh paragraph : 

In the same book, in the Story of Bambhadatta (Jacobi, p. 12, Meyer, p. 41) 
a Yaksa, gratified by a woman's devotion (bhatti = bhakli) brings about her 
marriage with king Bambhadatta. 

Page 27, after second paragraph, add : In Dandin's Dasakumaracarita Prince 
Arthapala becomes the husband of TaravalT, daughter of Manibhadra. 

Page 27, line 29, for " is Vasudeva," read : " are Siva and Vasudeva." 

Page 27, last line, add footnote to Kings : * All four are called Yakkhas in the 
Bharhut inscriptions. 

Page 28, line 3, insert : Mahavira is called Bhagavat, in the Aiipapatika Sutra, 

Page 30, note i, add: On Vajrapani see also Spooner, E. C, The Fravashi of 
Gautama, J. R. A. S., 1916. But Mrs. Spooner is quite wrong in thinking that 
"the conception of the guardian angel is un-Indian." Not to mention other 
examples we have the very word, arakkha-devata in Jataka, V, 429 and VI, 281 ; 
in the former case a father, in the latter a mother, reborn in the spirit world, i. e., 
as a Yaksa or YaksT, protects a child still living. An identification of Vajrapani 
with the Buddha's " external soul " is quite impossible. 

Page 30, the note beginning " Waddell " should be numbered "4" with 
reference to " Lag-na-rdo-rje." 


Page 31, note i, add: Excellent representations of Vajrapani attendant on 
Buddha will be found on the large slab from Nagarjunikonda, here plate 3, and 
on a slab illustrating the Marriage of Nanda, of the same Amaravati schools, 
both now in the Metropolitan Museum of Art, New York. 

Page 31, note 2, add: A more nearly complete Kusana statue of Vajrapani 
from Mathura is illustrated in Cunningham, Archaeological Survey Reports, III, 
pi. XI, d. 

Page 32, line 6. I use the designation Vrksaka (= Vrddhika, Varksi), of epic 
origin (Hopkins, Epic Mythology, p. 7) as most literally descriptive, and at 
the same time exactly equivalent to dryad. But Dr. Vogel ^ has recently shown 
that the word salabhailjika, " she who plucks sala flowers " was, or at least 
became, a technical term denoting representations of female types standing under 
trees, and was also the name of a ia/a-plucking festival in which women climbed 
the trees and plucked the flowers. Of course, it does not follow that the 
salabhanjikd in art always or even often stands for a human figure. There is 
every reason to suppose, that the Vrksaka is usually a Yakkhini; but Mahab- 
harata, III, 265, i-3a, "Who art thou that, bending down the branch of the 
kadamba tree, shinest lonely in the hermitage, sparkling like a flame of fire at 
night, shaken by the breeze, oh thou of fair brows ? Exceeding fair art thou, yet 
fearest naught here in the forest. Art thou a Devata, a YaksT, a Danavi, an 
Apsaras, or a fair Daitya girl, or a lovely maiden of the Naga king, or a Night- 
wanderer (RaksasI) in the wood " shows that the type could be identified in 
many ways. 

Evidently Yaksas, PI. 21, fig. 6, previously identified only as a scene from the 
Buddha's life illustrates the story of the Salabhanjika festival at Sravastl, Story 
No. 53 of the Avadana Sataka (Peer, p. 207, also cited by Vogel ; and my Notes 
sur la sculpture bouddhique, Rev. des Arts Asiatiques, V, 1928). 

Page 32, line 15, after "charm" add: cf. "the girdle of Aditi," IV. i, 5. 

Page 32, line 16, add : or fish-tailed elephants or horses ; the significance of 
these riverine vehicles will be discussed below, p. 47 f. 

Page 32, after line 22, add : A story is told in the Mahdbharata of a mother 
and daughter who embrace two trees, and thus become the mothers of Visvamitra 
and Jamadagni (Hopkins, Epic Mythology, p. 183). On trees granting offspring 
to women see also Bloch, Th., Notes on Bodhgaya, A. S. I., A. R., 1908-09, p. 142 ; 
Meyer, J. J., Sexual life in ancient India, pp. 156-8; and p. 12 below. 

In the Hatthipdla Jdtaka, No. 509, a poor woman has seven sons, and asked 
by whom she had them, she replies, pointing to the banyan tree by the city gate, 
" I offered prayer to the deity who inhabits this tree, and he answered me by 
giving these boys." The translations err in making the Rukkhadevata feminine. 

In the Dummedha Jdtaka, No. 50, people worship the Devata of a banyan tree 
( ? Kubera) for " sons and daughters, honor and wealth." 

In Bana's Kddambari, 134, Queen Vilasavati, desiring a child, performs a 
variety of ceremonies, amongst others " with a sunwise turn, she worshipped the 
pippala and other trees to which honour was wont to be shown." 

^ The woman and tree, or salabhanjikd, in Indian literature and art, Acta 
Orientalia, VII, 1928. 


In the Kathasaritsagara, Ch. CV, offerings of wine, flesh, and other dainties 
are to be made to the Yaksas on the wedding day. The bride here says to her 
husband, " Before I married you, I prayed to the Yaksas to enable me to obtain 
you " 

Page z^, Hne 2^, after "embrace," add: Atharva Veda, V, 5, 3 "Tree after 
tree thou cHmbest, Hke a lustful girl " ; and Jataka, V. 215, " When will Tarita's 
daughter .... cling to me, e'en as a forest-creeper to some forest-tree? " 

Page 33, note i, third paragraph, add: Cf. iipaddha sar'iram, describing the 
manifestation of a devatCi living in a royal umbrella {Jataka, VI, 376). 

Page 2)7, "ote i, add: Brown, W. N., The Indian and Christian miracles 
of U'alking on the zvatcr, Chicago, 1928. 

Page 38, line 8, image of " Manasa Devi": Mr. Cauda (A. S. I., A. R., 
1922-23, p. 165) reads the inscription as "(This image of the) Yaks! Layava 
has been caused to be made for the sons of Sii, and made by Naka, pupil of 
Kunika " ; in this case Kunika becomes the name of a sculptor, not of a Yaksa. 

Page 41, description of plate 16, 2: more probably Agni, cf. similar figure with 
ram vehicle on a pilaster of the Raj Rani temple, Bhuvanesvara. 

Page 42, plate 21, fig. 3. Now in the University Museum, Pennsylvania. 

Page 42, plate 21, fig. 6, read: The Salabhanjika festival at Sravasti, illus- 
trating the Avadana Sataka, Story No. 53; see Peer, L., Avadanagalaka, p. 207, 
and my Notes snr la sculpture bouddhiquc. Rev. des Arts Asiatiques, V, 1928. In 
Pancasikha's song in the Sakka-pauha Suttanta ( DN., II, 267 ^= Dialogues of 
the Buddha, II, 302) we have an allusion to the worship of the sal tree in full 
blossom " for offspring." For the fertility significance and pre-Buddhist char- 
acter of this festival see also J. B. O. R. S., XIV, p. 69. Vogel (The Woman and 
Tree or salabhaujikCi in Indian literature and art, Acta Orientalia, VII, 1929) has 
shown that the architectural term salabhanjika, applicable to " woman and tree " 
figures (e. gt. Yaksas, I, PI. 20, i ) is derived from the name of this festival. 

Page 42, plate 22, fig. i, for vrksaka read vrksaka. 

Page 42, description of pi. 22, 4, add : Height 3' 2}". Vogel, Catalogue 
of the Archaeological Museum, Mathura, p. 90. 

Page 43, Appendix, V, add : In Haribhadra's Sanaikiimaracarita, 514, Mayana's 
dyayana is likewise in a grove. In the Dasakumaracarita, Qi. V, the maidens of 
Avanti worship (arcayanti) Manobhava (= May ana, Kamadeva), laying their 
offerings of perfumes, flowers, turmeric, and Chinese silk on the level sand 
(saikata-iale) in the cool shade of a mango-tree. 


In Yaksas, I, and the addenda prefixed to the present discussion, 
there has been accumulated a mass of material sufficient to show the 
antiquity and popularity of the Yaksa cults in India, and to establish 
the general character of the Yaksa type, which includes universal dei- 
ties like Kubera, Kamadeva, and Sri, tutelary deities of kingdoms or 
clans, such as Satagiri and Sakya-vardhana, and also more localised 
and generally unnamed male tree spirits and dryads whose power does 
not extend beyond the shadow of the tree which is their abode. We 
have recognized that all these Yaksas, great or small, are vegetation 
spirits directly controlling, and bestowing upon their bhaktas, fertil- 
ity and wealth, or to use a single word, abundance. 

What we have not yet emphasized, though it has been indicated, is 
the intimate connection of the Yaksas with the waters. For example, 
Kubera's inexhaustible treasuries are a lotus and a conch, innumerable 
Yaksis have a inakara or other fish-tailed animal as their vehicle, 
Kamadeva the makara as his cognizance, the greater tutelary Yaksas 
control the rains essential to prosperity, and in the earliest mythology 
" that germ which the waters held first and in which all the gods 
exist " rose like a tree " from the navel of the unborn," who in the 
oldest passage is Varuna and in the Atharva-V eda is called a Yaksa; 
moreover, in the '' decorative " art,^ vegetation is represented indif- 
ferently as springing either ( i ) from the mouth or navel of a Yaksa, 
or (2) from the open jaws of a makara or other fish-tailed animal, 
or (3) from a "brimming vessel," or (4) from a conch, but never 
directly from any symbol representing the earth. 

Yaksas, then, are the Lords of Life, comparable to the Tuatha de 
Danann of Irish fairy mythology; they are also deities closely con- 
nected with the waters, though their habitat is terrestrial. These two 

^ Indian " decorative art " is not like modern ornament a kind of upholstery, 
but explicitly significant: as remarked by Kramrisch {Grundzuge der indischen 
Kunst, p. 83, " Das (indische) Muster ist so weit vom Ornament entfernt wie 
der Landschaft von naturalistischer Beschreibung." Even if the significance is 
partially forgotten in later periods, it is never wholly lost. Henceforward we 
shall usually speak of the decorative art by the names of its chief component 
parts, viz. the Animal Style, the Plant Style, and the Geometric Style. Each 
of these styles is a definite iconography, or at least tends to be recognizable as 
such to the extent that our knowledge advances. 



essentials of their nature are inseparably connected. A priori, it might 
have been supposed that the Nagas, who are water deities, and who 
control the activity of the waters, should have been the gods of abun- 
dance ; but they are not, as the Yaksas are, '* worshipped by those 
desiring children." The fact is, that the Yaksas control, not so much 
the waters as mere waters, but that essence {rasa) in the waters which 
is one with the sap in trees, with the auirta or elixir of the Devas, 
especially Agni, with the Soma, and with the seed in living beings. 

The Yaksa is by far a greater, more mysterious, and one may add 
potentially a spiritual power far more significant than the Naga or 
dragon.^ It is the object of the present work to discover (in the origi- 
nal meaning of the word) the importance of the Yaksas in what has 
often been vaguely referred to as a Life Cult, to suggest that this life 
cult, with which is also connected the worship of the Great Mother, 
may have been the primitive religion of India, and to show that the 
plant style is actually nothing more nor less than the iconography of 
the Water Cosmology. 

The term '' Water Cosmology " was first, I think, employed by 
Hume, in the Introduction to his Thirteen principal Upanishads, pp. 
10-14, with reference to such passages as Brhaddranyaka, 5, 5, " in 
the beginning this world was just water," and 3, 6, i, " all this world 
is woven, warp and woof, on water," and Kmisitaki, i, 7, where 
Brahman declares " the waters, verily, indeed, are my world." 

A belief in the origin of life in the waters was common to many 
ancient cultures, and must have arisen very naturally in the case of 
peoples, like those of the Nile, the Euphrates, or the Indus Valley, 
amongst whom water, in the form either of seasonal rains or of ever- 
flowing rivers was the most obvious prerequisite of vegetative in- 
crease ; nor can the belief be regarded as in any way unreasonable. 
Taken in a purely physical sense, it may indeed be called a fair an- 
ticipation of modern scientific ideas.^ In the Vedas, the belief appears 

^ Both are guardians of rain and wealth : but there seems to be indicated an 
original differentiation in this sense, that the Yaksas are bestowing, the Nagas 
withholding deities — perhaps a primitive dualism of good and evil. 

^A. Havelock Elhs, Studies in the psychology of sex, Vol. VII, Ch. VII, 
" Undinlsm." The rasa theory, one might say, presents a kind of analogy to 
our views about vitamines. Havelock Ellis, ib., 386, speaks of a " premonition 
of the modern scientific view of the pelagic origin of life ". While there is much 
that is suggestive in the whole chapter on " Undinism ", it may be remarked that 
scarcely anything in the Indian tradition seems to imply a connection of the 
cult of the waters and the vesical stream; while the origin of fountains, in the 
Mesopotamian tradition seems to be connected only with the symbol of the flow- 
ing vase (see Heuzy, Origines orientates de I'art, pp. 170, 171) but also Albright, 
Some cnices in the Langdon Epic, J. A. O. S., 39, 1919, esp. p. 70. Cf. Bolton, 


in the form of an old popular theory, for which are substituted the 
successively more philosophical concepts of a Space Cosmology, of a 
belief in an origin of the world in Non-being, in an origin of the 
world from Being, and finally in the conception of Brahman (the 
Absolute) as world-ground. The Water Cosmology, it is true, per- 
sists side by side with, and linked with these deeper views, even in 
post-Vedic literature; but it is not typically a creation of the Vedas, 
and seems to belong to an even older stratum of ideas than that which 
is developed in the Vedas. If the operations of the powers of vege- 
tative increase are not fully explained in the Vedas, it is because they 
belong to an older, pagan, fairy mythology, and the Vedas themselves 
seek to attain their ends rather by sacrifices (yajila) to celestial pow- 
ers than by the worship (piija) of localised personal and terrestrial 
deities or any attempt to stimulate their activity by suggestive repre- 
sentations in art."" The Water Cosmology conceives of certain powers 
of abundance who direct, or at least symbolise or represent the opera- 
tions of life as it wells upward from its source in the waters, and of 
a supreme deity, Varuiia ; but in itself, it can scarcely be regarded as 
a theology, for it does not originally conceive of a personal conscious- 
ness underlying the creation of the world. As remarked by Ronnow ^ 
" Der primitive Mensch denkt sich iiberhaupt keine konstante Per- 
sonification des Wassers ; es ist an und f iir sich heilig und von mana 
erfiillt " and precisely for this reason '' jedoch bald dieses, bald jenes 
Tier od. dgl. ein Repriisentant der dem Wasser innewohnenden Kraft 
werden kann." Thus a tendency to use abstract, non-anthropomorphic 
symbolism, which has so often been regarded as of distinctively north- 
ern, Aryan, or nomad origin, really inheres in the most ancient modes 
of thought. 

F. E., Hydro-psychoses, Am. Journ. Psychology, Jan. 1889. But a belief, 
erotically tinged, that water was the source of all things, has been at one time or 
another current in all the great civilisations, and traces of such a belief persist 
even in ^nodern European folklore. " The image of Aphrodite rising from the 
sea is not without scientific justification" (Donnan, in Ann. Rep. Smithsonian 
Institution, 1929, p. 318). 

* Or only in such cases as that of the pot dance of the maidens round the 
Marjaliya fire, after the horse sacrifice, for which an indigenous origin is 

Localised deities would not be expected amongst a nomad people, cf. Keith, 
Religion and philosophy of the Veda, p. 184, note 4. Further the natural em- 
phasis of nomadic people is laid on herds of cattle and on horses ; that of settled 
agricultural communities, probably already practicing irrigation long before 
the advent of the Aryans, upon water and plants. 

^ Trita Aptya, 1927, p. 6. 


There is nothing in this tendency contradictory to the use of the 
human form, which in the case of the feminine powers of fertihty 
and abundance can be traced far back into prehistoric times. These 
powers, particularly the nude goddess, the Great Mother, who may 
be Aditi, and Sri, who is so very closely connected with the Waters, 
stand in close relation to the Water Cosmology, and at the same time 
are represented in the aspect of women. But no stylistic distinction 
can be made as between this use of the human form, and the use of 
plant or animal forms ; the treatment is equally abstract and sym- 
bolic in both cases — there is no intention to reproduce natural appear- 
ances. A stylistic distinction, perhaps referable to diversity of ethnic 
preoccupation, may indeed be drawn between an expressive and an 
illusionistic art. But Indian art is never illusionistic ; it constitutes, 
in fact, a unity, and all the elements of this unity are congruous and 
coeval. Even to attempt to distinguish Aryan and Dravidian tenden- 
cies may represent in part a false issue. 

That detailed picture of the fairy powers at their procreative work, 
and of their hierarchy, culminating in the grandiose conception of the 
Regents of the Four Quarters, which the Vedas fail to give us, is to 
be found in the explicit formulae of the Plant Style, and incidentally, 
we may add, in Buddhist and Jaina literature, which, insofar as it 
refers to non-Buddhist beliefs at all, tells us much more about popular 
Indian religion than it does about the religion of the Brahmanical 
philosophers. At the same time, even the Vedic picture of the Water 
Cosmology and of the plants — " Plants, O ye mothers, I hail ye as 
goddesses," says the Yajur Veda (IV, 2, 6) — is more vivid and de- 
tailed than we might have expected. 

We are then in possession of two distinct sources, each equally per- 
meated through and through by the concepts of the Water Cosmology. 
One of these is the Vedas ; the other the Plant Style in the decorative 
art of the earliest monuments, and its later and even modern sur- 
vivals. These two sources are in complete accord. Hitherto, a con- 
centration of attention upon the sectarian, theistic art of India 
(Vaisnava, Bauddha, etc.) and upon stylistic development in con- 
nection with the cult image only, has obscured the fact that quantita- 
tively speaking, Indian art is to a greater extent than has been sup- 
posed, an illustration of Vedic ideas. We have already seen, for ex- 
ample, that in the representation of the abhiseka of Sri the elements 
of the dogmatic symbolism are far more ancient than the first extant 
representations.^ So now it will be found that the special formulae 

^ See my Early Indian iconography, II, Srl-Laksml, in Eastern Art, I, 1928, 
and Appendix, ch. II, 1929. 


of the Plant Style are only explicable in the light of the innumerable 
passages in Vedic literature in which the Water Cosmology is re- 
ferred to. It follows, of course, that the Plant Style did not come 
into being for the first time about 200 B. C. ; but that the Suhga re- 
liefs are simply the oldest monuments we possess of what is really a 
very ancient style. This is not surprising in itself ; it is merely a 
special case of the general argument for the long pre-Maurya an- 
tiquity of the earliest Indian animal, plant, geometric, and architec- 
tural formulae as met with in Maurya, Sunga, and later reliefs. 

Iranian cosmology, preserved in the Zendavesta, presents us with 
a body of belief and a type of gods and genii closely related to those 
of the Water Cosmology in India. Thus Ahura Mazda corresponds 
to Varuna ; Anahita and Ashi, his daughters, present a close analogy 
with Sri-Laksmi; the Amesha Spentas, especially Haurvatat and 
Ameretat, " Health " and *' Immortality," genii of plants and waters, 
have much in common with the Yaksas, and so in another way have 
the Fravashis, as " self-like genii " and guardian angels ; Apam 
Napat = Apam Napat ; haoma — soma; yasna = yajna. It is imma- 
terial for the moment that the ranks of Asura and Deva in Persia and 
India were reversed. The change took place in India, and Persia pre- 
served what can only be inferred in India from the oldest parts of the 
Vedas and from survivals in popular belief and art. Actually the 
Zendavesta gives us a better picture of Varuna than can be found in 
the Vedas themselves.^ 

But the cosmology presided over by Ahura Mazda is no more an 
entirely new creation of the Zend than the Water Cosmology is an 
invention of the Vedas. To complete the picture, indeed, we should 
have to go farther back, to parallels such as that of Ishtar with Aditi, 
and of Sumerian apsu (the underworld sea of sweet water) with 
Sanskrit a pair. As remarked by Masson-Oursel and demonstrated by 
recent excavation in the Panjab, " la solidarite indoiranienne, loin de 
se borner a une homogeneite aryenne prehistorique, est en fait quasi 
permanent a travers I'histoire, . . . . les rapports entre la Mesopo- 
tamie et le Pendjab ou la Serinde ont du etre frequents depuis 
Tantiquite sumero-dravidienne, et durant la conquete, jamais achevee 
de ITnde par les Aryas." ^ 

* Cf. Kretschmer, P., Varurm nnd die Urgeschichte der Inder, Wiener Zeit- 
schrift fiir den Kunde des Morgenlandes, XXXIII, 1926. 

^ Masson-Oursel, reviewing Abeg, E., Der Messiahglauhe in Indien iind 
Iran . . . . , in Journal Asiatique, CCXIII, 1928, p. 189. 

Cf. the remark by Przyluski, J., La vilte du Cakravartin: influences baby- 
loniennes sur la civilisation de Vlnde, Rocznik Orjentalistyczny, V, 1927, p. 21, 


It may be added that it is not without significance that the nearest 
analogies of Indian Siinga decorative art and architecture are not with 
contemporary Persian, but with Babylonian and other western Asi- 
atic, including Hittite, particularly from about the eighth century 
B. C. backwards. If borrowing in the Maurya or Suriga period is 
assumed, we should also have to postulate a selective archaism ! Then, 
too, of the most distinctive water-symbols, lotus, conch, makara, and 
inexhaustible vessel, the first three can only have become known to 
Aryans after their arrival in India while the last is a typically Baby- 
lonian as well as Indian conception. The elephant, a cloud and there- 
fore water symbol in connection with Indra and Sri is again neces- 
sarily of purely native origin ; the idea of guardian deities of the 
quarters is Babylonian as well as Indian.* 

Thus Indo-Iranian, and likewise Indo-European, cannot be re- 
garded, as they are commonly regarded, as altogether or even mainly 
synonymous with Indo-Aryan and Indo-Germanic. Indian and Euro- 
pean paganism. Life and Vegetation cults and the Fairy Mythology, 
have much in common that goes back to a late neolithic period. Indo- 
European would have had a meaning before the Greeks invaded 
Greece. A culture such as we might expect to have arisen in the later 
stone age and amongst permanently settled agricultural communities, 
and already embodying all the essential elements of civilization, ex- 
tended from northern India to Egypt and the Mediterranean as early 
as the third or fourth millenium B. C. Strzygowski's " Mazdean 
landscape " is probably Aryan only by inheritance. 

" les notions babyloniennes permettent d'interpreter, d'une fagon coherente, les 
faits indiens." Cf. Hertz, A., Die Kultur urn den Persischen Golf und ihre 
Aushreitiing, Leipzig, 1930; also, though to be read with caution, Hewitt, "It is 
in India that we find the original form of the religion which preceded that of the 
Semite-Accads in Assyria" {Early history of northern India, J. R. A. S., 1889, 

* For the cloud and mountain symbols see my A royal gesture, in the K. Bat. 
Genootschap Feestbundel, 1929, pt. I, and Notes on Indian coins and symbols, 
Ostasiatische Zeitschrift, N. F., IV, 1928. In architecture, compare the Indian 
volute capitals and pyramidal battlement forms with the similar forms at 
Khorsabad (palace of Sargon) ; the early Indian stupa with Phoenician and 
Chaldean tombs. Similarly in the case of the peculiar motifs of the Animal Style, 
particularly the designs of two or more animals with a single head common to 
all. Recent discoveries in the Indus Valley, proving at least a trade relationship 
between Indian and Mesopotamian centers in the third millenium B. C, point 
in the same direction. The early cult of a nude goddess, common to ancient India 
and Western Asia may be instanced (see my Archaic Indian terracottas, in 
Ipek, 1928). For the Hittite "Siva" see Garstang, The Hittite Empire, 1930, 
p. 20s, and cf. Ipek, 1930, pp. 53, 54. Cf. also, below p. 24, note 2, and p. 32. 


These matters are here touched upon, not because it is the primary- 
purpose of this work to prove the pre-Aryan origins of the Water 
Cosmology, but to make it quite clear that in using Vedic sources I 
am not asserting that its origins are in fact Aryan. Henceforth, the 
Vedic material will be drawn upon without regard to its mixed ori- 
gins; and it will be endeavored to show merely that a great part of 
later Indian art can only be understood in the light of ideas that are 
put forward more clearly and more constantly in the Vedas than any- 
where else — in other words, that the Plant Style is a survival of Vedic 
art, using the designation here rather in a secular than an ethnic sense. 

One further point is of considerable interest. In Semitic and Euro- 
pean conceptions of the Water of Life ^ the draught is conceived of 
as bestowing immortality forever. In India we meet with the more 
sober conception of repeated rejuvenation ; and this is equally true, 
whether we take the case of the gods whose life is renewed by repeated 
draughts of sonui, or that of human beings magically restored to life 
or rejuvenated by the good offices of Indra or the Asvins. All the 
life charms of the Atharva Veda are directed to restoration to health, 
or to longer or fuller life, never to immortality in a literal sense. And 
while in early India, and probably in a remoter past, all conceptions 
of well-being were thus connected with life on earth, and its perpetua- 
tion in offspring, the later development of philosophy altogether pre- 
cluded the possibility of the development of any theory of personal 
immortality, inasmuch as it was clearly realized that whatever comes 
into being must again disintegrate, and that only that can never die 
which has never come to birth. 


In Kadavul Mamunivar's Tiruvutavurar Purdna, Siva's immanent 
energy is compared to the heat latent in firewood; in the Bhagavad 
GUd, vii, 8 and xv, 13, Krishna says of himself " I am the vital 
essence (rasa) in the waters," and " It is I that as soma, very self 
of rasa, nourish all plants"; in the Lalita Vistara, vii, 91, we find 
" with the Water of Life (amrta) shalt thou heal the suffering due 
to the corruption of our mortal nature." Here in three of the later 
sectarian systems we find employed the language of an older mode of 
thought, adapted to theological or edifying purposes. We have al- 

^ Hopkins, E. W., The Fountain of Youth, J. A. O. S., XXVI, pp. 67 seq.; 
cf. Arbman, F., Tod und Unsterblichkeit im vedischer Glauhen, Archiv fiir Re- 
ligionswissenschaft, XXV, XXVI and Wensinck, A. J., The ocean in the litera- 
ture of the western Semites, 191 8, pp. 56 ff. 


ready cited pertinent passages from the Upanisads ; still older texts 
abound in the ideology of the Water Cosmology, and the best concep- 
tion of this ideology will be derived from the quotation of a series of 
typical passages, mainly from the Yajur Veda ' (YV) and the Atharva 
Veda (AV.). Thus : '' Those of which in the sky the gods make their 
food . . . . , those that inundate the earth with their rasa, the pure 
ones; may these waters be gentle and kindly to us " (YV., V, 6, i). 

" The plants born three generations before the gods .... The 
plants whose king is Soma, impel us to long life. Plants, O ye moth- 
ers, I hail you, O goddesses The fruitful, the fruitless, the 

flowering, the flowerless, impelled by Brhaspati, may they free us 

from harm Falling from the sky, the plants said * He whom 

we reach while in life, shall not come to hurt.' .... Food and 
strength do I take thence from the abode of holy order,'' from the 
birthplace of immortality. May it enter into us, in cattle and in plants ; 
I abandon decline, lack of food, and ill-health " (YV., IV, 2, 6 and 7). 

" Let flow the divine waters, the honey-sweet, for health, for 
progeny!" (SBr., VI, 4, 3.) 

" Let the heavenly waters, rich in milk {pdyas) flow propitious 
upon thee; propitious to thee be the herbs " (AV., VIII, 2, 14, 15). 

" Water, lightning, clouds, rain, let the liberal ones favor you. 
Anoint the earth, O Parjanya, with thy milk; by thee poured out, let 
abundant rain come " (AV., IV, 5, 6 and 9). 

" The waters divine do thou pour full of sweetness to avert diseases 
from men, from their place let arise plants with fair leaves " (YV., 
IV, 1,2). 

" The foetus of the waters and the plants is cattle * For 

the waters thee, for the plants I take,' therefore from the waters 
plants spring ' For the plants thee, for offspring I take, there- 
fore the food of man is plants ' . . . . therefore through Prajapati 
offspring are born " (YV., Ill, 3, 5, 6). " The plants are connected 
with Mitra, the waters with Varuna ; on the sap of the waters and of 
the plants do we live " (YV., II, i, 9). 

In connection with the first feeding of a child with solid food, we 
find : " I give thee to eat the essence of water and of the plants " 
(Hiranyakesin Grhya Sutra, II, i, 5). 

The waters used in royal consecration : "Ye ... . are the sap of 
the waters, of the plants .... the givers of the kingdom .... 
winning great radiance for the Ksatriya With the glory of 

^ I. e., Taittiriya Samhita of the Black Yajur Veda. 

' Rtasya: the reference can only be to the kingdom of Varuna, who is essen- 
tially the " Lord of rta." 


Soma I besprinkle thee (O king) .... to the son of the waters, 
hail!'* (YV., I, 8, ii, 12). Likewise "The waters of heaven that 
revel with milk, in the atmosphere and also on the earth — with the 
splendor of all those waters do I pour (abhi-sic) upon thee " (AV., 

IV, 7.5). 

"From rain originate virility, sap, well-being" (SB., I, 8, 3, 15). 

Soma : *' The soma .... indeed approaches the worshipper in 
the forn^ of Varuna " (YV., VI, i, 11). 

Soma in Vedic texts is often identified with the Moon, and like 
Soma, the Moon is often called the Lord of plants (cf. S. B. E., I, 
p. 286, note 2). 

The sonm of Tvastr obtained by Indra, is also called madhu, or 
mead ; its further identity with rasa, etc., is shown in the prayer ** That 
seminal fluid of ours, wondrous, abundant, may Tvastr release, as 
increase of wealth with good heroes, as offspring to us. O trees, let 
free . . . ." There are in fact many texts identifying the sofna with 
the essence in the waters, sap in trees, and seed in man and animals. 

" From trees is strength gathered ; the might of the waters sur- 
rounded by kine " (YV., IV, 6, 7) : " Of the waters the first-born 
rasa, likewise of the forest trees ; also Soma's brother art thou ; also 
virility of the stag art thou " ( (AV., IV, 4, i, 5) : '' They call Soma 
the seed of the strong horse " (YV., VII, 4, 19) : " I ask the seed of 
the male horse ; ^ I ask the navel of all existence .... this soma is 
the seed of the male horse ; this sacrifice is the navel of all existence " 
(AV., IX, 10, 13, 14; YV., VII, 4, 18). 

In AV., XIX, 31, 12, an amulet of udumbara wood is called virile 
(vrsan) ; ib,, XIX, 34, 8, the virtue of the forest tree (vanaspati) 
jangida is called its virility (virya), and this was bestowed on it by 
Indra of old ; ib., I, 35, 3, we have " the waters' brilliancy, light, force, 
and strength, also the virya of the forest trees." According to 
Mahabharata, I, 18, the amrta in the cosmic sea is derived from the 
sap of trees originally growing on Mt. Mandara, admingled with the 
waters in the process of churning. In the Mahabharata a. cycle is sug- 
gested ; we are told that a being fallmg from heaven to take new birth 

^This is in connection with the ritual with the mahisl in the Asvamedha, or 
horse sacrifice, the main purpose of which is to promote fertility. The sacrifice 
is to Prajapati, "the progenitor," but was probably originally to Varuna 
(Dumont, Asvamedha, p. xii; Eggeling in S. B. E., XLIV, pp. xix-xx; Johann- 
sen, Ueber die altindische Gottin Dhisand . . . . , 1917, p. 132). For a cosmic 
interpretation of the Asvamedha see the early part of the Brhadaranyaka 


becomes a subtle essence in the waters, and this water becomes 
semen ; * thence entering the womb it develops into visible life like 
fruit from flower; entering into trees, plants, air, earth, space, the 
same watery seed of life assumes the forms of quadrupeds and bipeds, 
and this is true for every visible creature. Cf. Chdndogya Upanisad, 
I, 1,2, *' The essence (rasa) of all beings is the earth, the essence of 
the earth is water, etc." 

Soma is in the milk of cows, for by eating and drinking the plants 
and waters they collect it (SBr., I, 6, 4, 4, 6). Similarly man '* Having 
collected that (Soma or moon) from the waters and plants, he causes 
them to be born from the oblations " (SBr., II, 4, 4, 20). 

The following is from the Hymn to the Honey-whip (AV., IX, I), 
recited when mixing soma with milk in the Agnistoma rite : ** Great, 
all-formed, the milk of it ; also they call thee ( Agni) the seed of ocean 
. . . breath of creatures, navel of immortality (ainrta). . . . Who 
knows that, who understands that which is the inexhaustible soma- 
holding vessel which is the heart of it? . . . its two unexhausted, 
thousand-streaming breasts, they milk out refreshment. . . . What 
honey on hills, on mountains, what in kine, in horses, in liquor {sura) 
as poured out, what honey is there, be that in me ! " The honey-whip 
(madhukasa) seems to be the lightning (Agni) that brings down the 
rains (cf. Indra's vajra) ; it is also personified as a goddess of abun- 
dance, presumably Aditi, since she is called '' the mother of the 
Adityas, the navel of amrta." As to the Asvins, some scholars regard 
them as Indo-Aryan, analogues of the Greek Dioskuroi, others as suc- 
couring deities of purely Indian origin : in Mahdhhdrata, I, 66, 40, the 
Asvins, plants and animals, are all called Guhyakas, and their chief is 
Kubera, which would make the Asvins out to be Yaksas. 

" In the sea is thy heart, within the waters ; let the plants and the 

waters enter thee I have penetrated to the waters, we are 

united with the rasa " (YV., I, 4, 45). 

*' O plants, do ye accept Agni here .... may he smite away from 
us misfortune. O plants, do ye rejoice in him, O ye that are rich in 
flowers, and have fair berries; this germ of yours .... hath sat 

him in his ancient seat Ye waters are healing, further us to 

strength, to see great joy ( = RV., X, 9, i, 3). The most auspicious 
rasa that is yours, accord to us here, like eager mothers " (YV., 
IV, I, 5). 

" In the waters, O Agni is thy seat, thou enterest the plants " (YV., 
IV, 2, 3). 

^ Cf. Aitareya Upanisad, I, 2, 3, " the waters became semen, and entered the 
virile member." 


When the clay for the Fire-pan is prepared: ** Thou (earth) art 
the back of the waters, expansive, wide, about to bear Agni " ; and 
again when the Fire Altar is prepared, the horse is led forward, and 
the lotus leaf laid on its footprint: "Thou (earth) art the back of 
the waters, the birthplace of Agni, the ocean swelling on either side ; 
growing to might as the lotus flower, do thou extend in width with 
the measure of heaven " (YV., IV, i, 3, and IV, 2, 8). Then when 
the gold disk (of the sun, the form of Agni in the sky) is placed on 
a lotus leaf "^ on the altar, the Satapatha Brahmana, VII, 4, i, 8, 
explains " The lotus means the waters, and this earth is a leaf thereof ; 
even as the lotus leaf here lies spread on the waters, so this earth lies 
spread on the waters," and ih., X, 5, 2, 8, " the lotus leaf is water." 
In the Jdiminiya Upanisad Brahmana I ,10, 2, the water cosmology 
is combined with the conception of the Absolute as world ground, 
thus " In it (the oik) the waters are established, in the waters the 
earth, in the earth these worlds." 

Closely connected with the Water Cosmology and with Yaksas is 
the idea of the productive pair, mithuna; the prominence of such pro- 
creative pairs in later art has been discussed by Gangoly,'' while in 
the earlier art such pairs are constantly recognizable as a Yaksa and 
Yaksi, and it may be remarked that the formula appears very com- 
monly in Sunga terracottas. The word mithuna is constantly used in 
connection with ritual coitus, e. g., that of the luahisi and the sacri- 
ficial horse (SBr., XIII, 5, 2, 2) and in connection with the Maha- 
vrata festival, the spring solstice, when the strengthening of the sun 
must have been the object in view (Ait. Ar., V, i, 5).^ These facts 
suggest the true explanation of the abundant representation of erotic 
pairs on the Sun temple at Konarak. 

The following passages are significant : ^' Mithuna means a produc- 
tive couple" (SBr., X, 5, 2, 8); "May I become born again, like 
kine; may I be glorious like a mithuna; mine be the msa Jn the 
waters, and the forms of the plants (osadhaya rupah) (Ait. Ar., V, 
I, i); "From Prajapati, when dismembered, couples (mithunani) 
went forth .... birth originates from a mithuna'' (SBr., IX, 4, 

'Cf. SBr., VII, 5, I, II, and VIII, 3, 2, 5, where avaka plants (regarded by 
Weber, Indische Studien, XIII, 205, as lotuses) are employed in analogous 
fashion and said to mean water. Similarly in YV., V, 4, 2. 

^ The mithuna in Indian art, Rupam, 22-23, 1927. See below, p. 33. 

^ For further references see Johannsen, Ueber die altindische Gottin Dhisand, 
p. 38, note I and p. 45. 


The navel : throughout the Vedas we meet with the phrase " navel 
of immortality (ainrta)'' in varied applications, of which several in- 
stances have been cited. The significance of the navel as a seat of the 
life-force is more precisely set forth in the Satapatha Brdhmana, V, 7, 
I, 9, where the sacrificer hangs a golden sun-disk around his neck so 
that it rests upon his navel, and it is explained " Why over the navel ? 
(Because) the navel is the seed, the power of procreation, and the gold 
plate represents vital energy and vigor " ; and in the Hiranyakcsin 
Grhya Sutra, I, 6, 24, i, " the navel is the center of the life-breaths 

In the Pauranik conception of the birth of Brahma, the creator is 
abjaja, lotus-born from the lotus that springs from Visnu's navel, 
said to represent the center of energy of the universe, while the lotus 
is the material aspect of evolution, the petals its consecutive forms 
(Agni Puraija, XLIX). Visnu, as Sayana-murti, here reclines upon 
the waters ; the great name Narayana is said to mean " moving on the 
waters " ; cf . Kubera's epithet Naravahana, iiara supposedly referring 
not to men (as later understood), but to water spirits, Gandharvas. 
That an expanded lotus represents the manifested universe (prapanca) 
is a commonplace of medieval symbolism. 

Although there does not seem to exist any representation of the 
birth of Brahma in sculpture dating before the sixth century A. D., 
the event is explicitly described in the MahdbJiarata (III, 2^2, 44, and 
XII, 207, 13). The former of the passages cited reads as follows: 
"As soon as that Eternal Being (Narayana) concentrated thought 
upon a New Creation of the Universe " a lotus immediately came into 
existence from His navel and the four- faced Brahma came forth from 
that navel-lotus." The extreme limits for the Mahdblulrata are from 

^ " The whole world, whatever there is, was created from and moves in 
prana" (Katha Upauisad, VI, 2); " prana is the life of all (sarzmyusa) " 
(Taittirlya Upanisad, II, 3). This importance attached to the vital center 
below the navel is illustrated in the art in the Bir'lh of Brahma formula, and that 
of the lotus rhizome rising from a Yaksa's navel : when the rhizome springs 
from a Yaksa's mouth, it is possible that the main (life-) breath, the Mukhya 
Prana, is also thought of as the source of the vegetative force (cf. Brhadar. Up., 
I, Z,7 and 8, where the Mukhya Prana is called the rasa of the limbs). But it is 
perhaps more likely that the saliva is regarded here as representing the waters. 
The word salila, applicable to saliva in particular as well as to water in general is 
used in Taittirlya Aranyaka, I, 23, i, as synonymous with dpah, the usual 
designation of " the Waters." For comparative matter on the navel see Wen- 
sinck, A. J., The ideas of the western Semites concerning the navel of the earth, 
1916; also above, p. 2 and below, p. 29. 

^ The Indian yuga system, here implied, has Sumerian sources or analogues, see 
J. A. O. S., 39, p. 66. 


400 B. C. to 400 A. D., but already in YV., IV, 6, 2 Visvakarnian 
( = Brahma) is born from the navel of the Unborn, in the waters, and 
cf. Br had Devatd, V, 154, 155. 

This tradition appears already in the RV., I, 24, 7, in connec- 
tion with Varuna, then in RV., X„ 82, 5, and YV., IV, 6, 2—'' Prior 
to the sky, prior to this earth, prior to the living gods, what is that 
germ which the waters held first and in which all the gods existed? 
The waters held that same germ in which all the gods exist or find 
themselves ; on the navel of the Unborn stood that in which all 
beings stood." The Unborn, of course, is one of the early designa- 
tions of the world-ground, later called Purusa, Prajapati, Brahman, 
or Narayana ; and with the full development of theism, Visnu inherits 
the formula. Meanwhile, in the AV., X, 7, 38, That One is spoken of 
as " a great Yaksa in the midst of the creation, lying upon the sea 
in penance ; therein are set whatever gods there are, like the branches 
of a tree round about a trunk." * In the Katha Upanisad, VI, i, we 
find " This eternal fig-tree! That (root) indeed is the Pure. That is 
Brahman. That indeed is called the undying. On it all the worlds 
do rest, and no one soever goes beyond it " ; and somewhat similarly 
in the Bhagavad Gitd, XV, 1-3. There is thus an ancient and con- 
tinuous tradition of a world origin in which are involved the waters, 
a Yaksa, the navel, and a tree of life, the latter first mentioned in 
connection with Varuna. 

It would be possible to multiply citations, of the kind assembled 
above, almost indefinitely. The nature of the Water Cosmology is, 
however, sufficiently revealed in what has been given. The ideology 
may be summed up as follows : from the primeval Waters arose the 
Plants, from Plants all other beings, in particular the gods, men, and 
cattle. Rasa, as an essence of the Waters, or as sap in trees, is vari- 
ously identified with soma, amrta, semen, milk, rain, honey, mead 
(fiiadhu) and liquor (surd) ; there is a cycle in which the vital energy 
passes from heaven through the waters, plants, cattle and other typi- 
cally virile or productive animals, and man, thence ultimately return- 
ing to the waters. The clouds rain milk or soma; they are sometimes 
called cows, as is also Aditi, the goddess of abundance who is also 
a personification of the honey-whip of the Asvins, which may be the 
lightning. The myth of actual creation takes the form of the origina- 
tion of a tree from the navel of a Primal Male, who rests upon the 

* Many scholars have preferred in this passage to translate " Yaksa " simply 
as a "great wonder." See above, p. 2, where I have argued that the original 
word should be retained. 


Waters, and from whose navel the tree rises up ; he is called a Yaksa 
and was originally Varuna. 

The abode or source of Agni is in the waters, in plants (wood), or 
in the earth, as well as in the sun and lightning. The prayers ad- 
dressed to the Waters, or to the Plants, or to deities controlling them 
or other members of the series, are all of the nature of instigations to 
function vigorously. The lotus generally is a symbol of the waters, 
the lotus leaf which lies on the back of the waters is specifically a 
symbol of the earth ; the waters are the support of all things. 

In Vedic ritual there is an extensive use of vessels of water, often 
brought into connection with Varuna, and this has survived up to 
the present day." 


Varuna and Mitra in RV., VII, 65, 2, are asurd dryd, " noble 
asuras," and even in AV., i, 10, Varuna is still an astir a who rules 
over the gods and whose commands are fulfilled ; cf . Jdimimya Brdh- 
inana, III, 152. 

An antithesis of Devas and Asuras in the Vedas has long been 
recognized, and it has been held by many scholars that the Asura 
gods, of whom the chief is Varuna, belong to another family, known 
in India long before that whose chief is Indra. Thus Charpentier 
remarks in connection with the Vedic mythology that " While gods 
like Indra .... seem to be the lords of a rustic, semi-nomadic, 
strong and half -barbarous generation, Varuna and Mitra seem to be 

in close touch with a much higher civilization If Indra is the 

somewhat grotesque chief of a flock of early knights-errant, Varuna 
is the king in a well-ordered city-state .... it seems to be a more 
or less unavoidable conclusion that these gods were once introduced 
amongst the Indo-Iranians from some other people." ^ 

The dominant theme of the Vedas is that of the conflict between 
the Devas and Asuras {e. g., RV., I, 108, 6 and X, 124 ; YV., V, 4, i). 

^ Cf. Burgess, J., The ritual of Ramesvaram, Indian Antiquary, XII, 1883. 

^ Charpentier, reviewing Keith's Religion and philosophy of the Vedas, in 
Bull. Sch. Oriental Studies, IV, p. 339. Cf. Ronnow, Trita Aptya, p. 75: ** es 
steht mit ziemlicher Sicherheit fest, das nicht die Devas .... die ursprung- 
lichen Besitzer der Soma v^aren, sondern eben die Damonenwesen." " Demons," 
in such cases, generally represent the deities of an older and subsequently re- 
jected mythology, as in the case of the European fairies, cf. Alfred Nutt, The 
fairy mythology of Shakespeare, London, 1900. 

Cf. Oldenberg, Religion des Veda^, pp. 187 ff., where it is suggested that Va- 
runa, the Sun, and other Adityas were not originally Indo-European gods, but 
were taken over by the still united Indo-Iranian Aryans from Semitic {so. 
Sumerian) sources. For the moral contrast between Varuna and Indra cf. also 
Giinters, Der arise he Weltkonig und Hciland, p. 97. 


Finally " The gods drove out the Asuras, their rivals and enemies from 
this world " (SBr., XIII, 8, 2, i). Varuna, indeed, escapes this fate, 
for he is accepted as a Deva,^ and his asuric character is forgotten; 
but other Devas, Indra, Prajapati, Narayana, inherit his high func- 
tions, and he, as Varuna specifically, is reduced to the level of a god 
of the sea and of the waters generally, a sort of Indian Neptune, but 
with many reminiscences of his original character. 

Probably the best discussion of Varuna will be found in Kret- 
schmer, P., Varuna unci die Urgeschichte der Inder, WZKM., 33, 
1926. The connection of Varuna with Greek 'Ouranos is now mis- 
doubted. Kretschmer sees in the Vedic Varuna a combination of a 
Hittite sea god, Aruna, with the Indo-Iranian Asura, Iranian Ahura 
Mazda ; i. e., finds that the Indian Varuna embodies two elements, an 
Iranian (Aryan) and a Sumero-Accadian-PIittite, the latter due to 
borrowing or inheritance by the " ur-Inder " (Aryan Mitanni) from 
the Sumero-Accadian culture surviving in Mesopotamia. 

It would appear to me, however, that as god of living waters, fer- 
tility, and justice, and as a great king, Varuna belongs almost en- 
tirely to a settled order of things, to a city state and peasant culture 
of immemorial antiquity ; that on the dark chthonic side of things, 
with its seasonal festivals, ritual eroticism, and possibly human sacri- 
fice, the whole complex of ideas connected with Varuna and Aditi, 
Gandharvas, Yaksas, and so forth, points backward to a great culture 
evolved with the beginnings of agriculture, and flourishing from the 
Mediterranean to the Indus, rather than to the priestly invention of 
later warlike peoples, such as the Persian or Indian Aryans. Varuna 
and Aditi in many respects suggest Tammuz and Ishtar.'' 

It is as king that Varuna's noose or fetters (pdsa) are called into 
play as penalties for sin. These fetters are drought, and the disease 
yaksma, perhaps dropsy. Prayers and offerings are constantly ad- 
dressed to Varuna, for release from these fetters. Sometimes other 
deities are asked to release the rivers or to absolve from sin, thus 
Agni and Soma " freed the streams from the dread imprecation, when 
they were held fast by Varuna's fetter" (YV., II, 3, 14). In Hille- 
brandt's view the Agnistoma of the spring festival is offered to 
Varuna for the release of the rivers from their winter fetters. The 

^ Cf. Brown, W. N., Proselytising the Asuras, J. A. O. S., 39, 1919. 

^ For Varuna's fundamentally chthonic character and his relation to Aditi as 
mate and son, cf. Johannsen, K. F., Ueher die altitidische Gottin Dhisand . . . . , 
pp. 132, 133; for Aditi, cf. AV., VII, 6, i, and Jdiminiya Upanisad Brdhmana, 


scapegoat sacrifice at the end of the horse sacrifice, when a man repre- 
senting Varuna stands in water and receives the sins of the com- 
munity upon his head is suggestive in this connection. Johannsen, 
pp. 125 fT., sees here a survival of human sacrifice to Varuna, or 
rather, of Varuna himself ; he thinks that a Purusamedha preceded the 
Asvamedha, and like Hillebrandt sees in both the survival of the 
sacrifice, as a vegetation ritual, of a temporary king, for which so 
many parallels have been adduced by Frazer in the Golden Bough; 
the transition from a fertility to a sin offering in the case of the 
scapegoat ritual, being a later development. There is certainly suf- 
ficient evidence for a practise of human sacrifice to trees (tree 
spirits) in early India. 

The ideal of kingship embodied in the original conception of 
Varuna may be said to have persisted in Indian culture up to the 
present day; it is very evident in the person of Rama. The ideal king 
is a Dharmaraja, an incarnation of justice, and the fertility and 
prosperity of the country depend upon the king's virtue ; the direct 
connection between justice and rainfall here involved is highly sig- 
nificant. Some more special points may be briefly noticed ; thus, in 
Iranian mythology, earthly kingship (divine right) is plainly estab- 
lished and dependent upon a kingly glory, hvarena, " made by Ahura 
Mazda," and overshadowing every legitimate king. The idea is rather 
less prominent in India, though yams, royal glory, and tejas, fiery 
brilliance, partly correspond in usage to hvarena. The idea appears, 
however, in a more specific form in Java and Cambodia, and though 
in connection with Siva, rather than Varuna himself, embodies many 
ancient features ; for though the Devaraja or deified principle of 
kingship is here represented in the form of a lingam, this is a fiery 
emblem, and the setting up of such a lingam marks the establishment 
of a hegemony and secures the prosperity of the kingdom/ The simi- 
larity of the lustration (abhiseka) of a king in the coronation cere- 
mony, and the abhiseka of Sri will not be overlooked (cf. above, 
p. 21) ; and the connection of royalty with rainfall will be found again 
in connection with a characteristic gesture of a Cakravartin as repre- 
sented in early reliefs." Finally, may not the superiority of the 
Ksatriya to the Brahman in Buddhism and Jainism (systems devel- 
oped notoriously in incompletely Brahmanised areas, and often pre- 

^ Bosch, F. D. K., Het Lingga-heiligdom van Dinaja, K. Bat. Genootschap 
Kunsten en Wetenschapen, LXIV, 1924, esp. p. 272. Cf. below, pp. 43 ff . 

^ Coomaraswamy, A. K., A royal gesture, in Feestbundel K. Bat. Genootschap 
Kunsten en Wetenschapen, Deel I, 1929. 


serving popular non-Vedic features, especially the Yaksa cult) repre- 
sent a survival from a time, equally pre-Aryan in India and in 
Persia, when kingship implied divinity and ranked above priesthood ? 

The character of the Vedic and Epic Varuna as summarised in 
Macdonell, Vedic mythology and Hopkins, Epic mythology, may 
be taken for granted as known to the reader. We shall discuss here 
only such points as have the most direct bearing on the present prob- 
lem. Most prominent in the personality of Varuna are his connec- 
tion with the celestial waters, and with holy order (rta) physical and 
moral; his kingship {kmtra, samrdj) and justice, and the fetters 
(pdsa) with which he binds the sinner and controls the waters. At 
first sight, the logical connection between these qualities may not be 
obvious ; but actually it is one that has remained prominent through- 
out the history of India. It is precisely upon the virtue and justice 
of any earthly king that the falling of the rains and ripening of the 
crops in due season directly depend; when a king's virtue fails, the 
order of Nature is disturbed. There is an ordeal by water (in which 
Varuna is specially mentioned), oaths are taken upon water, the 
bride circumambulates fire and water ; a lying or even careless wit- 
ness '' casteth a thousand pdsas of Varuna upon himself" (Mahd- 
hhdrata). As suggested by Professor Brown,' it is most likely Varuna 
as keeper of the waters and guardian of truth, who makes the "Act 
of Truth " efficacious ; and who is the witness of the sealing of a gift 
or contract by libation (daksinoda) . 

Varuna was originally the root of the Tree of Life, the source of 
all creation (RV, I, 24, 7), and it is presumably still Varuna who is 
called the Unborn in RV., I, 24, 7, Unborn and " the Recumbent " 
(uttdnapad, with legs outstretched) in RV., X, y2, 2 and 3, and a great 
Yaksa reclining in tapas upon the back of the waters in AV., X, 7, 38, 
where the tree springs from his navel ; though this formula is soon 
inherited by Prajapati (YV, V, 6, 4), and then by Narayana (Visnu) 
who retains it to this day (see above, pp. 2, 3). The world tree as 
species is variously interpreted in the literature, most often as an un- 
dying asvattha or nyagrodha^ as mmhina in the Suparnddhydya, and so 
forth, but as represented in the Plant Style and in connection with the 
Birth of Brahma, as a lotus.^ 

^ Indian and Christian miracles of Walking on the Water, p. 9. 

^ The nyagrodha is called " Varuna's " in Gobhila, Grhya Sutra, IV, 7, 24. 
Later, e. g. lataka no. 489, the banyan is particularly connected with Kubera, and 
various unnamed rukkha-devatas. 

^ For further details, see pp. 2, 3, 13, and 24 above, and my Tree of Jesse and 
Indian parallels or sources, Art Bulletin, XI, 1929. 


As lord of holy order, the succession of the seasons belongs to 
Varuna's lordship, and there is good reason to suppose that the great 
seasonal festivals, as suggested for example by Hillebrandt for the 
Agnistoma of the Spring Festival, offered for relief from the fetters 
of winter, were primarily offered to him/ In Mbh., IX, 50, 32, those 
who perform the caturnulsya and the one hundred and ten sacrifices 
go to the " abode of King Varuna " ; the Varunapraghasas, the rites 
of the second of these four-month oft'erings, are for the remission of 
sin, by confession and offering to Varuna. 

There is ample reason to believe that the soma offering and the 
horse sacrifice were originally made to Varuna, and only later trans- 
ferred to Indra and Prajapati." For the soma, some texts have been 
cited above, and in RV., V, 85, 2, we have " Varuna has placed 
Agni in the waters, the Sun in heaven, and Soma on the rock " ; ib., 
IX, 95, 4, Varuna is clearly a synonym for soma; ib., X, 31, 6, 
Varuna is called " the wise guardian of the ainrta "; and SBr., IV, 
I, 4, 9, soma cva Varunasya; but the general argument is even more 
cogent than any selected text. The hurse sacrifice is a vegetative ritual 
designed to secure the establishment of sovereign power, the fertility 
of men and cattle, and absolution from sin. Amongst evidences of its 
certainly pre-Vedic and probably pre-Aryan antiquity is the fact that 
certain characteristic features, such as the intimacy of the MahisT 
with the slain horse (the pair is designated a mitliuna), and the ob- 
scene dialogue are somewhat reluctantly tolerated rather than in- 
vented by the Brahman authors of the ritual texts.' The original con- 
nection with Varuna is preserved in the statement of the sacrificer, 
" He who will kill the horse attacks Varuna " (Asvalayana, Sraiita 
Sutra, X, 6, 10, and SBr., XIII, 4, 3, 5), and in the ritual of the final 
bath. It is not easy to see why the horse should be associated with 
Indra; but a natural connection of the horse with Varuna and with 
the waters is stated or implied in many places, e. g., SBr., V, 3, i, 5, 
'' the horse is Varuna's own," and ib., VI, 2, i, 5, where the horse is 

^Hillebrandt, Vedic Mythologie, II, p. 40. 

'^Eggeling, in SBE., XLIV, xviii-xxiv; Keith, Taittiriya Samhita, HOS., 
Vol. 18, pp. cxxxiv-vii, and Rel. and phil. of the Veda, HOS., Vol. 32, p. 346; 
Dumont, L'Asvamedha, p. xii ; Hillebrandt, Ved. Myth., II, pp. z:^ ff. 

^ On maithnna cf. above p. 23 ; Gangoly, O. C, The mithuna in Indian art, Ru- 
pam, 22-23, 1925; Mukherji, B. L., in Woodroffe, Sir John, Shakti and Shakta, 
pp. 441, 442. In SBr., ix, 4, i, 2, Gandharvas and Apsarases are said to have pro- 
ceeded "in pairs" from Prajapati {sc. Varuna) and "birth originates from a 
pair.'* In this auspicious progenitive significance of " pairs " lies the explanation 
of the constant representation of mithuna, and sometimes of mdithuna in the later 
art, as at Konarak. Cf. the bhutdnam mdithuna at the Mahdvrata, where the pur- 
pose is to strengthen the sun in its northward course. 


slaughtered " for Varuna," while in YV., II, 3, 12, 2, horses are 
apsu-yoni, '' water-born " and the white horse produced at the Churn- 
ing of the Ocean provides a specific example/ In the Brhadaranyaka 
Upanisad, I, 2, 7, the Asvamedha horse is indeed said to be Praja- 
pati's; but Prajapati's connection with the ritual is of course a very 
late development, and this is by no means the only case in which he 
inherits what properly belongs to Varuna, whom, indeed, he represents. 

In the Mahdhharata, Vana Parva, Varuna bestows on Bhrgu a 
thousand horses which arose from the Ganges; in SBr., X, 6, 4, i, 
the sacrificial horse is identified with the whole universe, and the pas- 
sage concludes " The sea, indeed, is its kindred, the sea its birth- 
place." The existing iconography does not show us horses in con- 
nection with Varuna, but both normal and fish-tailed or water-horses 
(jala-turaga) are commonly found as vehicles of Yaksas and YaksTs. 
The wide distribution of the idea of water or sea horses (e. g., Russia, 
Greece, Scotland, China, etc.) is an indication of its antiquity. 

It would appear natural to connect the doctrine of the Lokapalas, 
the Four Maharajas, Regents of the Quarters, with the old descrip- 
tions of Varuna and Ahura Mazda as " four-cornered " (RV., I, 
152, 2, and Vendidad, I, 18) ; for though it does not occur in Indian 
literature before the Yajur Veda,' the Four Guardians and the World 

^ Cf . Keith, loc. cit.; Charpentier, Suparnasage, p. 385; Johannsen, Uehcr die 
altindische Gottin Dhisand . . . . , pp. 132, 151-3 (The horse is Varuna). 

Although the Churning of the Ocean (samudranianthana) , a myth of the 
creation of the sun(-horse) and moon (soma, amrta), etc., and a "Water Cos- 
mology " myth of the first importance, is only described at length in the Epic, 
and only represented in Gupta and later sculpture, it is plainly referred to (as 
pointed out by Charpentier, Suparnasage, pp. 383 ff.) in RV., X, 72, and this 
fact and the parallel myths in other countries show its remote antiquity. Inci- 
dentally, as remarked by Charpentier, it should be noted that in RV., X, 72, 6, 
iva nrtayatam should be rendered, not as by von Schroeder in Mysterium tind 
Mimus " as dancers," but " as if dancers," or " like dancers," the clouds of spray 
raised by the gods (who are not dancing, but churning) being compared to 
clouds of dust raised by the feet of dancers. Cf. RV., X, 82, 6, and YV., IV, 6, 2. 
" This germ the waters first bore, when all the gods came together," and RV., 
I, 163, 1-4, and YV., IV, 6, 7, "Arising from the ocean or the spray .... like 
Varuna to me thou appearest, O steed." 

^ The earliest assignments of deities to the four quarters are those of YV., 
I, 8, 7, where we get Agni (E), Yama (S), Savitr (W), and Varuna (N), 
Brhaspati (Zenith), and ih., VI, i, 5, where we find Pathya Svasti (E), Agni 
(S), Soma (W), Savitr (N), and Aditi (Zenith); ib., II, 4, 14, Indra is 
guardian of the East. In AV., I, 3 the immortal guardians are praised, but not 
named. The Sadvimsa Brahmana, LV, 4 and SBr., Ill, 6, 4 have Agni (E), 
Yama (S), Varuna (W), Soma (N) ; and other schemes occur, those of the 
Buddhists and Jainas differing, usually with Kubera in the North. 


Mountain actually offer a very striking parallel to what is found in 
the Babylonian systems, and they have been thought to have been 
borrowed, or as I should express it, inherited from such sources ; ^ 
it is certainly not inconceivable that a cosmology cognate to the Chal- 
dean may have been known in India in pre-Vedic times. In the Brah- 
manical and Jaina systems, Varuna himself is Regent of the West, 
but it is more natural and accordant with his original status as 
supreme ruler to think of the four regent kings as his vassals" con- 
ceived in his likeness, and acting as his delegates, somewhat in the 
manner of the Amesha Spentas ; in the Buddhist system, which as 
usual seems to embody older and more popular ideas, the Four Re- 
gents, viz., Vessavana = Kubera (N), Dhatarattha (E), Virulhaka 
(S), and Virupakkha (W) have as their subjects Yaksas, Kumbhan- 
das, Gandharvas, Apsarases, and Nagas, all beings connected with the 
waters, and in the Bharhut inscriptions, accompanying their images, 
all four are called Yakkhas. In all the systems, the Four Great Kings 
have space elephants (dig-gajas) as their vehicles, and possibly these 
elephants themselves were the original guardians of the quarters ; in 
any case, they are the sources of the winds, which they blow through 
their trunks, and in the ahJiiscka of SrI-LaksmI two or four of them 
pour down the rains from the inverted cask or jar of the clouds, which 
vessel in RV., V, 85, 3, 4, is specifically Varuna's. It is true that in 
the later mythology Airavata, the chief of the dig-gajas, is specifi- 
cally the vehicle of Indra ; but this is not Vedic, nor is it in the least 
degree likely that a purely Aryan deity should originally have been 
associated with a distinctively Indian animal symbol. In all proba- 
bility the elephant, like the horse, was an ancient symbolic element in 
the Water Cosmology ; for though we do not find it in iconography 
directly connected with Varuna, we do find normal sky- or cloud - 
elephants {dig-gajas) and water-elephants (jalahasfin, jalchha) asso- 
ciated with Sri-LaksmI, with Yaksas and Yaksis as vehicles {Yaksas, 
I, pi. 3, fig. 2 ; pi. 4, fig. 2 ; pi. 22, fig. I ; and here pi. 9, fig. i), and in 
the I^lant Style in connection with loti^s vegetation (pi. 11, fig. i, and 
40, fig. 2), and iv is further noteworthy that the dwarf Yaksas (pi. 43, 
fig. 7, also the merman, pi. 41, fig. i) are often elephant-eared, and 
this suggests a connection with Ganesa.' 

^ Kirfel, W., Die Kosmographie der Inder, pp. 28-36, esp. p. 34*. Cf. Heine- 
Geldern, Weltbild und Banform in Sudostasians, Wiener Beitrage, IV, 1913; 
Warren, W. F., Problems still unsolved in Indo-Aryan cosmology, J. A. O. S., 
XXVI, p. 84; Przyluski, J., La ville du Cakravartin, loc. cit. supra. 

^ Cf. the Digvyasthapana, the " mounting of the king on the quarters," a part 
of the Rajasuya ceremony. 

^See Yaksas, I, p. 7, and pi. 22,, fig. i, and MFA Bulletin, No. 159. 


In SBr., XIII, 4, 3, 7-8, King Varuna's people are said to be 
Gandharvas, and those of King Soma, Apsarases ; these, are closely 
associated divinities of the waters and of fertility, and originally of 
more significance than when in the later literature they become little 
more than the musicians and dancers of Indra's court. The Gand- 
harva or Gandharvas seem to have been original guardians of the 
soma on behalf of Varuna (cf . " Gandharvas, overseers of the guard- 
ians of the soma," SBr., Ill, 3, 3, 11, Kanva recension), and this 
is clearly why the soma for the ^-o/z/a-sacrifice when made to Indra 
has to be purchased from the Gandharva (Ait. Br., I, 2y, i), and why 
Indra is generally in the RV. hostile to the Gandharvas (RV., VIII, 
I, II, and 66, 5). The Gandharva Krsanu, who clearly corresponds 
to the Avestan Gandarewa Keresani who is connected with the haoma, 
is called a soma-pdla in Ait. Br., Ill, 26, 3, 2, and is said to be an 
archer and to shoot at the eagle which carries off the soma (RV., 
IV, 27, 3, and Tait. Ar., I, 9, i) ; but in the Badami reliefs," follow- 
ing a later version of the story, it is Varuna himself who, seated on 
his makara, shoots at Garuda. Amongst the defenders of the soma 
in the epic version of the story are both Yaksas and Nagas." Refer- 
ence has already been made to Gandharvas and Apsarases as tree 
and fertility spirits (Yaksas, I, pp. 32, 33)." The importance of Gand- 
harvas and Apsarases as progenitive deities appears also not only in 
their connection with marriage, but also in SBr., IX, 4, i, 2 and 4, 
where they are said to be produced in couples (mithuna) from Pra- 
japati, for "birth originates from a pair (inifhtma)." Apsarases are 
sometimes swan-maidens, swimming in lakes in the form of water 
birds (SBr., XI, 5, 1,4, in connection with the story of Urvasi) ; per- 
haps some reminiscence of this idea ought to be recognized in the con- 
stant representation of hamsas, amongst or perched on expanded lotus 
flowers, in the Plant Style (cf. PL 11, fig. 2, center) ; it may also be 
"significant that they are constantly represented as carrying lotus 

^Banerji, R. D., Bas-reliefs of Badanfi, Mem. A. S. I., 25, pi. XXIII, e. For 
good discussions of the story see Bloomfield, M., The legend of Soma and the 
Eagle, J. A. O. S., XVI, 1894, and Charpentier, Suparnasage. 

^ Nagas are not generally closely connected with the soma, but are indicated 
amongst .yo77w-guardians by the epithet "footless " in SBr., I, 7, i, i, and Siipar- 
nadhyaya, 22, i, and 23, i. In the early iconography, both Yaksas and Nagas 
may carry amrta-Hasks ; the former are notably addicted to intoxicating liquor. 

^ " Nyagrodha, Udumbara, Asvattha, Plaksa .... are the homes of the 
Gandharvas and Apsarases" (YV., Ill, 4, 8.) As to how and why a connection 
between trees and human fertility may first have been imagined, cf. Przyluski, J., 
Totemisme et vegetalisme dans I'Inde, Rev. de I'Histoire des Religions, XCVI, 
1927, p. 359- The sequence of vegetable propagation is easily observed ; and at a 


flowers or garlands in their beaks. To sum up, Gandharvas and 
Apsarases appear to have been at first genii of vegetation and fer- 
tility, connected with Varuna and Soma, and when later they are 
reduced to the status of attendants on Indra, they are replaced, func- 
tionally, by the Yaksas and Yaksls. Yaksas and Yaksis are identical 
with Gandharvas and Apsarases as originally conceived, and perhaps 
this is a point on which considerable emphasis should be laid, as 
partially explaining some of the numerous other links which seem 
to connect the Yaksas, including King Kubera, with Varuna. 

The intimate connection of Varuna with Soma, and the partial 
and early identification of both with the Moon are noteworthy, as is 
also the fact that in the later art, in Navagraha groups, the Moon is 
sometimes given the makara-vdhana which is properly Varuna's. 
There does not seem to exist any very early source for the association 
of Varuna with the makara, although the latter is a very obvious 
symbol of the waters ; but as ^-o/zza-guardian in the Mahdbhdrata ver- 
sion of the Rape of the Soma and in the corresponding Badami re- 
liefs he is shown seated on the makara, beside the soma which is 
represented in the form of a jar placed on a rock, and again similarly 
seated, but letting fly an arrow at Garuda."" 

Few myths are recorded in connection with Varuna; but there is a 
suggestion in AV., IV, 4, where his virility is decayed and is restored 
by means of a herb ' dug for him by a Gandharva, that the very im- 
portant theme of the god whose potency is impaired, with disastrous 
cosmic results, may once have belonged to Varuna. A hint of the 
usual explanation of the weakness appears in the Rdnmyana, VII, 56, 
12 ff., where Varuna begets Vasistha upon UrvasI, " who belonged to 
Mitra, but loved Varuna " ; but this may nevertheless be an ancient 
legend. There is also the Epic story of Varuna's theft of Bhadra, 
daughter of Soma (the Moon) ; Bhadra's husband, the sage Utathya, 
punishes the god by drinking up the waters, so that the land became a 
salt desert, a sand-waste. Only when the " Water-King " surrenders 

time when paternity was not yet understood, the fruiting of trees and the growth 
of seedlings provided primitive man with an apparently obvious explanation of 
the nature of human conception. The idea of conception by the eating of a fruit, 
still current in Indian folklore, presents a phase of this idea in which the repro- 
ductive potency of a tree or tree spirit is evidently the supposed active agency. 
Cf. Meyer, J. J,, Sexual life in Ancient India, pp. 156-8, and 561. 

^ Still functioning as soma-guard'ians in Kansitaki Brdhmana, 12, 3. 

^Banerji, loc. cit., pi. XXIII, a and e; here, pi. 45, fig. i, and see p. 30. 

^ The plant is called " the first-born rasa of the waters and also of the plants, 
brother of soma and the lusty force of the antelope buck." 


Bhadra does Utathya release the waters and set the world free from 
affliction/ Somewhat in the same way Soma, the Moon, for his uxori- 
ousness towards RohinT, is punished by the curse of Daksa, who brings 
yaksma upon him, which results in a waning which devastated the 
world and frightened even the gods ; nor can this be wondered at, 
since the Moon *' is water-born soma, without which nothing is pro- 
duced " (Mbh., XIII, 67, II fif.) ; the only cure that can be found is 
in recourse to the ''six essences of Varuna " (Mbh,, I, 66, 17, and 
IX, 35, 43 f.).^ It is almost needless to repeat that Varuna, Soma, 
and the Moon are constantly identified. These myths are more fully 
developed in connection with Indra, Prajapati, Agni and Siva, but 
there is at least a suggestion that Varuna may have been the original 
" Fisher King." The problem is further discussed below, p. 37 f¥. 

Another ancient Indian deity, who seems to have belonged to a 
mythological cycle outside the range of the Vedic tribes, and is con- 
nected with the soma, is Tvastr, whose *' mead " (niadhn) is called 
sometimes the '* food of the Asuras," sometimes the " food of the 
Devas." He is said to have fashioned for the gods a special cup, but 
it is significant that the drink has to be stolen from him by Indra. In 
RV., V, 42, 13, an incest of Tvastr with his daughter is suggested. 

We have already (above, p. 2) traced a continuity of the myth 
of the world tree springing from the navel of the cosmic deity, reclin- 
ing on the Waters, from Varuna to Visnu. Finally, the striking re- 
semblance in ethical character between Visnu and Varuna may be 
remarked ; Visnu in heaven, Rama on earth, are both ideally righteous 
and wealthy kings, on whom as such naturally depend the prosperity 
of the universe or the earth. It is also noteworthy that Visnu's com- 
plexion is said to be blue like the ether, while Varuna's is said (Vi- 
snudharmottara) to be of the hue of water when the sky is reflected 

^ Cf. Meyer, J. J., Sexual life in Ancient India, II, p. 318. 

^ An earlier, almost identical version of the same story is found in YV., II, 3, 5 
and II, 5, 6, 4-5 where Soma has to wife the thirty-three daughters of Prajapati, 
but favors only Rohini ; he suffers from yaksma, hence called " king's evil " ; 
and is cured by the new moon oblation to the Adityas, which makes him wax. 

The text adds that he who knows the origin of these ailments will not be 
visited by them. Everyone will be familiar with the Brahmana and Aupanisadic 
tendency to stress the importance of knowledge about a ritual, even above its 
performance ; and with the universal Indian view that as from pupil to teacher 
only the asking of the right question can provoke the right answer. There may 
be a parallel here to the primary importance attached in the Grail Quest to 
the asking of the right questions ; in India, magical efficacy is attributed to a 
statement of the truth. 


in it/ In the Agni Piirana, Ch. LXIV, the identity of Varuna and 
Visnu is actually recognized. A connection of Visnu's consort LaksmI 
with Aditi has been suggested.^ 

It is thus practically proven, though only a part of the evidence has 
been presented above, that we have to do with, not a succession of new 
cosmic and supreme deities, but with a succession of new names at- 
tached to the original conception of the cosmic deity, the succession 
being that of Varuna, followed by Prajapati, Purusa or Brahman, 
(Svayambhu), and Narayana or Visnu. This is further supported by 
the fact that all of these, taken two or three at a time, are in one place 
or another, of the literature, explicitly identified. The typically inter- 
mediate name Prajapati, *' the Progenitor," is indeed an epithet rather 
than a name, and as suggested by Johannsen,^ may always be taken 
as directly the equivalent of Varuna. A recognition of these facts is 
of fundamental importance for the interpretation of Vedic and Hindu 

In other words, the creation myths of the water cosmology (espe- 
cially the Churning of the Ocean, and the World-tree myth in its 
various forms), which are later so conspicuously connected with 
Visnu, are really inherited from Varuna. In the same way a succes- 
sion of designations of the great Mother and Earth goddess can be 
recognized in Aditi, Ida, Dhisana, Prakrti, Vak, and Laksmi and 
Bhumi Devi, and in all aspects of the concept of Sakti. 

The description of Varuna in the VisnudJiarmoftara, III, 52, though 
late, is not without interest and significance. He rides in a chariot 
drawn by seven hamsas, said to represent the Seven Seas, he has an 
umbrella of dominion, and is supported by a makara. He has some- 
what of a hanging belly (like a Yaksa : cf. AV., IV, 16, 3, " Varuna's 
paunches " and ih., IX, 15, *' a paunch (iidara) for treasure ") ; he is 
four-handed (this is of course a post- Vedic development shared with 
other deities), holding the lotus and fetter (pdsa) in his right hands, 
conch and jewel-vessel {ratna-pdtra) in the left. The conch is said 

^ Cf . Jdiminlya Upanisad Brahmana, IV, i, i, where Varuna is called 

^ Cf. Eastern Art, I, 1928, p. 175 ; also Varunani = Laksmi (Monier, Williams, 
Skt. Diet.). 

""Ucber die altindische Gottin Dhisana . . . . , p. 132, note: "Prajapati ist 
der brahmanische nachfolger Varuna's, ist ein ander, ein noa-name, der an stelle 
der tabuierten names Varuna getreten ist." For the identity of Daksa, Purusa, 
and Prajapati, see Charpentier, Suparnasage, p. 391, discussing RV., X, 90. 

Cf. also the connection of sura (Varuni) with Varuna, later with Prajapati 
(Hopkins, E. W., The Fountain of Youth, J. A. O. S., XXVI, p. 67.) 


to represent riches (cf. Kubera's sankha-mdhi) ^ the fetter to bind the 
samsdra, the umbrella to be glory (yasas, cf. below, pp. 28, 45, and 
AV,. VI, 39, where the prayer for yasas is a prayer for sovranty), the 
makara well-being, enjoyment, or fertility (saukhya). His wife is 
Gauri (in the Rdmdyana, Gauri or Varum), holding a blue lotus in 
her left hand. Attendant are Gahga on the right, holding a lotus and 
standing on a makara, said to represent virility {inrya), and Yamuna 
on the left, holding a blue lotus and standing on a tortoise (cf. p. 53), 
said to represent time (kdla).^ It will be seen that Varuna's original 
character as a great king, dispenser of justice and punisher of sin, lord 
of rivers and of increase, is well preserved, and that the concrete 
symbolism is consistently and satisfactorily explained. 


The essential features of the Grail legend of Western Europe are 
the existence of a land ruled by a great king, the " Fisher King," 
whose land and castle are by the sea ; upon his vitality the prosperity 
and fertility of the country depend ; but notwithstanding that he pos- 
sesses an all-wish-granting talisman (the Grail itself), often described 
as an inexhaustible bowl or dish, but sometimes as a gem, he lies 
wounded '* in the loins " and impotent, or apparently dead, and his 
country is a waste land, parched by drought, and barren. The Grail 
quest is achieved when the hero, visiting the castle of the Fisher King 
and witnessing the ritual of the Grail procession and other marvels, 
enquires their meaning ; immediately the wounded king is restored to 
vigor, the rivers once more flow in their channels, and the land is 

The three generally current " Grail theories " are respectively the 
Christian, the Folk-lore, and the Ritual. The latter is the most satis- 
factory, and seems to be that accepted by a majority of Grail scholars. 
According to this view, the essential elements of the Grail legend, 
apart from the later Christianising, are derived from an ancient life 
and vegetation cult, ultimately perhaps of Western Asiatic or even 
Indian origin. In its ritual aspect, the Adonis cult provides the near- 
est parallels, while the fundamental theme of the Freeing of the 
Waters is typically developed, as we have already seen, in the Vedas. 

^ Cf . Varum sankha of the Epic (Hopkins, Epic mythology, pp. 116, 117). 

'Cf. "the fetter of time and the fetter of Varuna " (Mbh, XII, 227, 82 
and III) : and AV., XIX, 53, 3, where a full vessel is set on time. 

^ The right question would provoke the right answer : and this right answer 
would have the efficacy of an " act of Truth." 


Historically, these motifs found their way into European tradition 
as mysteries in the Roman period, and seem in later times to have 
been the leading ideas in a heretical Christian order probably to be 
associated with the Knights Templars. The medieval Grail literature, 
particularly in its earlier forms, embodies very many Oriental fea- 
tures, accessory to those of the main theme/ 

There is nothing novel in the recognition of Grail parallels in In- 
dian literature.' They are, however, more striking and more numer- 
ous than Grail scholars have suspected, and it will be useful to cite 
the most important. It may indeed be possible to indicate the outlines 
of a Life Myth connected with King Varuna ; and behind him there 
may lie some even more ancient Iranian or Indo-Iranian pre-Aryan 
Lord of Life ; for the concept of a Life deity (Tammuz) upon whose 
vitality the very existence of Nature and all its reproductive energies 
depended, and who was yet himself subject to declining powers and to 
injury or death like an ordinary mortal, was already a crystallised for- 
mula expressed in ritual observances in Sumeria early in the third 
millennium B. C. This Tammuz appears to have been not merely a 
vegetation spirit, but as suggested by Langdon, originally to have 
represented the vivifying waters, and like Varuna he was called a 
" son (or son-consort) of the waters." ^ 

It should also be borne in mind that perhaps the " mysteries " of 
the life-cults had always an esoteric as well as an exoteric aspect, as 
we know to have been the case in the immediately pre- and post- 
Christian period in the Mediterranean area where they were '' con- 
sidered not only the most potent factors for assuring the material 
prosperity of the land and folk, but were also held to be the most 
appropriate vehicle for imparting the highest religious teaching." " We 

^ It will suffice for present purposes to cite J. L. Weston, The Quest of the 
Holy Grail, 191 3, and From ritual to romance, 1920, where the subject is treated 
at length from the point of view of the Grail student, and where further refer- 
ences may be found. 

^ See, for example, Weston, loc. cit.; von Schroeder, Die Wurneln d. Sage v. 
heiligen Gral, Wiener Sitzungsberichte, Phil.-Hist. Kl., Bd. 166 and Arische 
Religion, Vol. 11 (see index) ; Meyer, J. J., Sexual life in ancient India, p. 400, 
note 2. 

^ Langdon, S., Tammuz and Ishtar, p. 5. Cf. Barton, G. A., in J. A. O. S., 
XLV, 1925, p. 35, "Tammuz of the Deep." 

* Weston, From ritual to romance, p. 149, commenting on the Refutation of 
the Christian mystic Hippolytus, ca. 228 A. D., where the mysteries are traced 
back to Assyrian sources on the one hand and on the other said to be fulfilled 
in Christianity. The phrase of Hippolytus " these Naassenes frequent what are 
called the mysteries of the Great Mother, believing that they obtain the clear- 
est view of the Universal Mystery from the things done in them " could be 
exactly applied to the followers of modern Indian Sakta cults. 


have already seen reason to believe that in India the deepest aspects 
of rehgious experience and the elements of metaphysics seem to have 
been connected rather with non-Vedic than with Vedic elements in 
Hinduism, and there can be no doubt of this so far as religious ecstasy 
is concerned ; the explanatory tendencies of the later Vedic literature, 
and the constant readiness of the Upanisads to draw a parallel between 
macrocosm and microcosm, may well represent rather the emergence 
of old traditions than an actual novelty, and it would be reasonable 
enough to suppose that it had always been understood that generation 
is an image of regeneration. However these things may be (and it 
should not be forgotten that Hindus have always believed and still 
believe in the great antiquity of the more profound ideas embodied 
in such systems as that of the Saiva Siddhanta), it is of interest to 
observe that just as in Europe the Grail legend motifs, originally 
pagan, were ultimately interpreted in an edifying and Christian sense 
(though never with the full approval of the Church), so in India the 
phraseology and symbolism of the life cults were retained and reinter- 
preted in sectarian circles (cf. p. 19 above), and in connection with 
deities other than those with whom they were first connected. 

Thus, in the Manimekhalai ' (a south Indian Buddhist legend dating 
in its literary form from about the third century A. D.) it is foretold 
of the heroine Manimekhalai that '* there will appear a damsel with 
a begging bowl (originally the Buddha's) in her hand. Fed from that 
inexhaustible bowl the whole living world will revive. As a result 
of her grace, rains will pour in plenty at the command of Indra, and 
many other miracles will take place in this town. Even when rains 
fail, the country will still have abundance of water." The bowl itself 
is called Amrta Surabhi and it appears once a year on the Buddha's 
birthday, from the waters of a lake beside a miraculous Buddha-seat 
protected by Indra; it emerges from these waters and enters Mani- 
mekhalai's hands ; she makes it her vocation to alleviate hunger, thus, 
for example, in Puhar " she appeared in the hall of the hungry and 
destitute, with the inexhaustible bowl in her hand, as if pouring rain 
had come on a wild region burnt up with the heat of the sun,^' and 
from the bowl she feeds all men to their uttermost satisfaction. The 
story is long and intricate, but it may be observed that the heroine is 
the daughter of Kovalan and Kannaki (the hero and heroine of an- 
other Tamil poem, the Silappadhikdram) , the latter being identified 
with the goddess Pattinl, extensively worshipped in Ceylon; that 

^ Vinson, J., Legendes houddhistes et djainas, Paris, 1920 ; Aiyangar, S. K., 
Manimekhalai in its historical setting, Madras (1928) ; Pope, g. v., in Siddhanta 
DIpika, Vol. XI, XII. 


Manimekhalai is protected by and closely associated and ultimately to 
be identified with a goddess Manimekhala/ who causes the destruction 
of a city by a tidal wave ; in fact, the whole story is packed with 
miraculous features, which are merely made into the means of edifica- 
tion from a Buddhist point of view. 

From such a reworking of ancient material as this let us turn to 
follow up some of the older sources. It will be found at once that 
almost every important Indian deity is said, in one place or another, 
to possess a wish-granting talisman, either an inexhaustible bowl or 
productive jewel, or a tree of paradise that yields all kinds of trea- 
sures, or a wishing-cow, or some other treasure, for example the 
sahkha and padma nidhis of Kubera ; in the Mahahhdrata, the Sun- 
god gives to KuntI a copper dish of inexhaustible food (von Schroeder 
recognized here a Grail motif) ; in the Manikantha Jdtaka (No. 253) 
a Nagaraja possess a precious gem which yields '' rich food and 
plenty " at will ; Kubera, in the Epic, is said to possess a *' beloved 
thing," which " gives immortality to mortals, makes the blind see, 
and restores youth to the old " ; it is kept in a jar guarded by 
dragons, in a cave very difficult of access. We have further the 
general and very significant fact of the drink or food of the gods 
{soma, anirta, etc.) always conceived as contained in or drunk from 
a special vessel, c. g., the cup fashioned for the gods by Tvastr. 
When the soma is represented in art, it is as a full vessel (pi. 45, 
fig. i), and precisely such a full or brimming vessel {punnaghata, 
etc.) is the commonest of all Indian symbols of plenty (see pp. 61 ff), 
and also, as a symbol of the waters in the Plant Style, is con- 
stantly represented as a source of vegetation. One may also mention 
the cup, cask or udder of the clouds, originally Varuna's, later Indra's, 
from which the rain or heavenly soma is poured down to earth, and 
in the ahliiscka of SrI-Laksmi is held inverted by the dig-gajas, or 
cloud elephants. Varuna himself as lord of waters carries a vessel 
and is called the Lord of Vessels. Many of the deities carry an amrta 
flask in their hands ; this is particularly the case with Indra, who often 
uses the contents to restore the dead to life.' The full vessels regu- 
larly carried by river goddesses (nadi-devatds, see pp. 66 ff), who can 
fairly be called Apsarases in the original sense of the word (water- 

^ See Levi, S., Manimckhala, a divinity of the sea, Indian Historical Quar- 
terly, VI, 1930. A " virgin of the seas " probably pre- Aryan, is still wor- 
shipped by fishermen on the Coromandel coast (Siddhanta Dipika, XII, p. 169). 

^ When an a?wrfa-flask is later carried by Avalokitesvara, no doubt the living 
water is spiritually interpreted, cf. Yaksas, I, p. 31. 


nymphs, cf. the Indian derivation apsu-rasa) may also be noted (pis. 
19, fig. 2 and 26, fig. i), and likewise the universal Indian custom of 
offering a full vessel to an honored deity or guest (see p. 6i, note 2, and 
pi. 41, fig. 4). 

At this point further attention must be called to one of the most 
characteristic features of the Grail legend and of Indian culture, 
though the idea is widespread elsewhere, viz., the direct connection 
between the virtue (moral and physical) of the king, and the fertility 
(dependent on rainfall) of the country over which he rules. This 
motif is so constantly met with in Indian literature at all periods, that 
it will suffice to cite a single typical example from the already quoted 
Manimekhalai, where a goddess addresses a prince as follows : " Oh, 
son of the great king! If the king swerve ever so little from right- 
eousness, the planets themselves will desert their orbits ; if the planets 
change their course, rainfall will diminish ; with a shortage of rain- 
fall, all life on earth will cease; the king will often cease to be re- 
garded as king, because he would seem not to regard all life as his 
own." We have seen already in Varuna the ideal prototype of the 
righteous, justice-dispensing, king, who makes the rains fall and the 
rivers flow, and so bestows fruit fulness upon the whole world : that 
some of his functions are later taken over by others is immaterial. 
We have also seen that the great deity possesses a vessel containing 
the Water of Life. 

Under normal circumstances it would appear that the possession 
of this Water of Life ensured the renewal of the vitality of the deity 
who year after year with the return of Spring, restored the world to 
life and productivity, after a season of apparent impotence or im- 
mobility; it is with the same end of restoring or maintaining the vital- 
ity of the god that the Vedic soma-o^trmg is made. Nevertheless, 
there must have arisen from time to time occasions of unparalleled 
and unseasonable drought and famine, which could not be sufficiently 
accounted for by the failings of an earthly king, nor remedied by 
human offerings or penance. It was surely natural to assume that 
such disasters were due to an impotence of the Divine King, and that 
this impotence or maiming must have been a consequence of some 
heinous sexual sin, either grave adultery or incest, for which the 
earthly punishment would be extirpation of the male organ.^ The 
necessity of immediately restoring the Divine King's virility would be 

^In SBr., II, 5, 2, 20, adultery is called a sin against Varuna, and must be 
expiated by confession (truth). The punishment mentioned above is that pre- 
scribed in the Dharma-sastras, e. g., Narada-smrti, XII, 75. 


obvious, for even the gods are appalled at the results/ The process of 
restoring the Divine King to vigor, though imagined to have taken 
place in heaven, would naturally be enacted as a drought-dispeUing 
ritual on earth ; actually most of the Vedic rituals in whole or part have 
the intention of restoring or increasing the power of the gods, or of 
their representative on earth, and we need only suppose a more special 
case, to have before us a simple and adequate explanation for the 
development of a *' Grail ritual." "^ 

We have already seen how in such a predicament Varuna is cured 
by a herb identified with rasa and soma, and how the Moon is reme- 
died by the six rasas of Varuna himself. These are perhaps versions 
or inversions of one and the same story, which later on we find again 
in connection with Prajapati, who replaces Varuna as a progenitive 
deity. The story ^ is best preserved in Ait. Br., Ill, 33-34, and SBr., 

I, 7, 4 f . : Prajapati in the form of a buck {rsya) couples with his 
daughter in the form of a deer (rohit) ; the gods are shocked, and 
invoke a dread form of Rudra, who wounds Prajapati with an arrow, 
so that his seed falls to the ground ; the gods, however, are not willing 
that it should be wasted, and after treating it with fire, fanned by the 
Maruts, various beings, animals, etc., are produced ; Prajapati be- 
comes the constellation Mrgaslrsa (" Deer's head ") and his daughter 
the asterism RohinT, whom we have already seen as the too much 
favored wife of the Moon. A somewhat similar story is more briefly 
indicated in connection with Daksa, in RV., X, ^2 ; here Daksa is the 
male principle ; and Aditi, at once his mother and daughter, and she 
becomes the mother of the gods, '' the friends of the amrtaf Daksa 
can be identified with Purusa (RV., X, 90) and with Prajapati (SBr., 

II, 4, 4, 2), and presents analogies with Varuna;" a reminiscence of 
the story as told of Prajapati can be recognized even in the later 

^ For the importance attached to the king's virility, cf. SBr., IX, 4, i, 4, "He 
alone is (ruler) of kingdom who propagates offspring." 

^ Cf. YV., I, 4, 45, " O bath, O flood . . . . , thou hast removed by sacrifice 
the sin committed by the gods" and ib., V, 3, 12, "He who sacrifices with the 
horse-sacrifice makes Prajapati whole; verily be becomes whole; this is the 
atonement for everything, and the cure for everything. All evil by it the gods 

^EarHer allusions in RV., I, 71, 5 and X, 61, 5-7 (Dyaus taking the place of 
Prajapati). A later reminiscence apparently in Mbh., I, 118, where a rsi's son 
is shot by Pandu while coupling with his mate in the form of a deer though 
here it is Pandu that is regarded as the sinner, and suffers from the curse that 
any indulgence in sexual intercourse on his part shall result in his immediate 

* Cf. Charpentier, Suparnasage, p. 391 ; and above, p. 36. 


legend of Daksa's sacrifice, where Daksa's head is replaced by that 
of a buck/ Somewhat similar stories are told of Indra in post-Vedic 
literature, where he is called a habitual paramour ; he is unmanned, 
by Agni's advice, for adultery with Ahalya, the wife of the sage 
Gautama, and for other sins, not sexual, he is several times paralysed. 
An earHer amatory escapade of Indra alluded to in RV., I, 51, 13, is 
explained in the Sadvimsa Brdhmana as referring to the seduction of 
Vrsanasva's daughter, the Apsaras Menaka (Oertel, H., Contribu- 
tions from the J aiminlya Brdlimana .... , J. A. O. S., XXVI, 176). 
Now in the soma sacrifice, the purchase of the soma by the gods from 
the Gandharva(s) in exchange for Vak '' because the Gandharva is 
fond of women" (SBr., Ill, 2, 4) forms the theme of a kind of 
ritual drama in which a Sudra represents the Gandharva (YV., I, 2, 7 
and SBr., Ill, 3, 3, 10; Caland and Henry, U Agnistoma, p. 46; Hille- 
brandt, Ved. Myth., 2nd ed., I, pp. 257-8) ; and it is most significant, 
in view of the fact that the offering is primarily to Indra, that the 
purchased soma is placed by the priest on the sacrificer's bared thigh, 
with the formula " Enter the right thigh of Indra," and that the 
sacrificer then rises, saying '' With new life, with good Hfe, am I 
risen after the immortals." Indra also loses his energy as a conse- 
quence of his struggle with Vrtra, the demon of drought ; his power 
and strength went into the earth and became plants and roots, and this 
is why soma is in the milk of cows, for plants are their food ( YV., II, 
5, 3; SBr., I, 6, 4, 4-6; Tdittiriya Samhita, II, 5, 3, 2 seq.) ; his 
strength is restored by soma. The stories as related of Varuna (see 
p. 34 above). Soma, Tvastr, Prajapati, and Daksa seem to be all 
forms of one and the same myth ; in the case of Indra, the stories 
are perhaps more trivial, but it is still significant that we have to do 
with a god of rain injured '' in the loins " or paralysed as punishment 
for a sin. It may be observed that the incest had not perhaps been 
originally so much regarded as a moral sin, as an infringement of a 
tabu (in any case some kind of incest on the part of the first progeni- 
tors is more or less inevitable, e. g., in the case of Manu and Ida in 
the flood legend, SBr., 1,8, i ) ; the result is a particular f ruitf ulness, 
but still the penalty of the infringement cannot be avoided. 

The somewhat more elaborate myth of the same kind connected 
with Siva^ may be still another version of the same legend, or an 

* For comments on the story in connection with Daksa, see Charpentier, 
Suparnasage, pp. 390-392. 

^ See Jahn, W., Die Legende vom Devadanwana, ZDMG., 69, 1915, and Die 
Legende vom Devaddruvana im Siva-Purdna, ih., 71, 1917; Deussen, W., Ueber 
das Devadaruvana, ib.; Bosch, F. D. K., Het Linga-Heiligdom van Dinaja, 
Tijdschr. K. Bat. Genootschap Kunsten en Wetenschapen, LXIV, 1924. 


independent story derived from the same stratum of ideas. Siva is 
said in RV., X, 92, 5, to release the waters, and the later Gahgava- 
tarana legend presents the same idea in a more extended myth ; but 
the Devadaruvana legend, though it is found only in post-Vedic 
works, mainly in the Mahabharata and in many of the Puranas is in 
its extant form, as pointed out by Deussen, a legend designed to 
make orthodox what was once an ancient cult of fertility : specifi- 
cally to explain and justify the lihgam cult.' The story, which has 
numerous variants, and is often represented in Indian sculpture ' may 
be summarised as follows : there is in the Devadaruvana forest a her- 
mitage of rsis, fire-worshipping sages, living with their wives and 
daughters, and practising rigorous asceticism. Siva visits the wood 
in the form of a naked mendicant, carrying a skull-cup and begging 
alms. At the sight of his indescribable beauty, the wives of the sages 
are distracted ; they follow him, casting off their clothes and orna- 
ments, and yield to his embrace. The enraged sages launch a curse 
at the mendicant, so that his liugani falls to the ground ; it strikes the 
earth, splits it open, and sinks into the underworld. The body of the 
mendicant himself falls into the opening. Dread portents follow, and 

^ As to the nature of the original fertility symbolism, cf., Przyluski, J., Non- 
Aryan loans in Indo-Aryan, in Bagchi, P. C, Prc-Aryan and pre-Dravidian in 
India, 1929, pp. 10-15, where it is suggested that linga, lanyala, and laiigula, all 
having amongst other meanings that of mcmhrum virile, and the second mean- 
ing also more usually " plough," are derived from a common non-Aryan root 
having the general sense of " to push in, to make a hole " ; and that the use of a 
planting stick, or in later cultural development, of the plough, was thought of 
as a fertilising penetration of Mother Earth, analogous to ordinary sexual inter- 
course. Cf. the production of Sita from the furrow with a plough by king 
Janaka ("the progenitor") ; YV., IV, 2, 5, and 6, where seed is sown in the 
field-womb, and the propitious plough " ploughs up a cow, a fat blooming maid," 
etc. ; also the importance attached to the ritual ploughing, by the ruler in person, 
at the beginning of the planting season in various Oriental countries, e. g., Ceylon 
and China; an Indian ploughing festival {vappamangala) of this kind is described 
in Jdtaka, I, 57 and DhA., II, 113. An analogous symbolism is to be found also 
in SBr., I, 9, 2, 21, where the contact of the male vcda (broom) with the female 
vedi (altar) is said to effect "a union productive of offspring." Cf. Sophocles, 
Antigone, 569. 

A curious and rather different account of the origin of Uiigam worship ap- 
pears in Mbh. X, 17, 8 ff. ; here Siva pulls off his own member, which drives 
into the earth and there stands erect. 

"^ E. g., relief at the Mallikarjuna temple, Srisaila, Karnul District (Longhurst, 
A. H., in A. S. I., A. R., Southern Circle, 1917-18, p. 32, and PI. XIV, b) ; a re- 
lief at Karusa (Burgess, Antiquities of Bidar and Aurangabad, A. S. W. L, 
vol. Ill, PI. XVII, 4) ; and several unpublished, amongst others, pillar figures in 
the entrance hall of the MTnaksT temple at Madura. 


the cosmic order is disrupted. The gods in terror hastily repair to 
Brahma, who explains that this is the result of the maiming and dis- 
appearance of Siva (until now the mendicant has not been recog- 
nized). The gods then resort to Siva, who lies swooning as if in deep 
sleep, and beseech him to resume his lingain, lest the three worlds 
perish. After a preliminary refusal, Siva agrees, upon condition that 
gods and Brahman s shall forthwith worship the lihgam. The gods 
agree to this, and the lingain is worshipped in the underworld. Siva 
is satisfied, and taking the Ungam, sets it up in the *' Field of the Lord 
of Gold " (Hatakesvara-ksetra, ? land of Kubera). At the same time 
Brahma sets up a golden lingain, called Lord of Gold (Hatakesvara), 
and proclaims that all who worship such a Ungam made of precious 
substances shall attain the highest path. As shown by Bosch, relying 
mainly on the Prahia Upanisad, III, 5, 8, and 9, Siva (Rudra) is here 
clearly the earthly form of Agni, hence the earth can be regarded as 
his body, and setting up of the Ungam in the earth efifectively accom- 
plishes the reunion of the member with the body of the god ; which 
is not without a parallel in the Grail ritual task of the " welding of 
the sword." 

Agni himself is an Asura king and is sometimes identified with 
Varuna or Mitra-Varuna (Macdonell, Vedic mythology ; Hopkins, 
Epic mythology, pp. 178, 222, 227, etc.), and also with Siva. The 
Devadaruvana legend actually occurs in a confused form in both Epics 
in connection with Agni (see Hopkins, Epic mythology, p. 194), who 
desires the wives of the rsis, and in order to seduce them assumes the 
form of the household fires ; unsuccessful, he seeks to " commit suicide 
in the forest " ; but Svaha, daughter of Daksa, loves Agni, and assumes 
the forms of the wives, and this results in the birth of Skanda. Here 
Svaha, daughter of Daksa, and regularly recognized as the wife of 
Agni, is clearly the same as Uma, daughter of Daksa and lover and 
wife of Siva, in the story of Daksa's sacrifice and the birth of Skanda 
(the War god). At the same time we have here, in still another 
variant, the story of the deity who sufTers for his transgressions ; it 
can hardly be doubted that Agni's wish to commit suicide is simply 
an expression of loss of power, the dying down of the flame, while 
the end of the story proves that virility has been restored. 

^ Dr. Bosch shows further that the story provides an adequate explanation of 
the "fiery lingam " of the Devaraja cult of Java and Cambodia, where it repre- 
sents the " fiery essence of kingship, a radiant earthly emanation of royal wis- 
dom and dominion"; for the male organ is the tejas of the lower Hfe breath 
(apdna) which corresponds to the earthly form of the cosmic deity, and this 
tejas (or yasas) may be regarded as cognate to the Iranian hvarena. Cf. above, 
pp. 28, 2>7' 


It is thus beyond question that the fundamental theme of the Grail 
legend is present in India, that it once belonged to a vegetation or 
fertility cult, and that just as in Europe, so in India the original mean- 
ing of the motif was gradually forgotten, so that the myth became a 
tale, employed for edifying purposes remote from those of its primary 

A few other parallels may be more briefly noted. The place of the 
Doctor in European vegetation rituals has been remarked upon by 
Miss Weston (Ritual to romance, Ch. VIII) and attention has been 
called to the medical value of herbs as stressed in these rituals and 
in the Vedas, esp. RV., X, 97, and also to the activities of the Asvins 
in Vedic myths. Varuna himself is a patron of physicians. In this 
connection there might well have been mentioned Dhanvantari, the 
Divine Physician who at the Churning of the Ocean, rises from the 
waters with the desired soma, accompanied by the apsarases (Rdmd- 
yana, I, 45, 20).' The fish symbol and the designation Fisher King 
have also been discussed, with reference to Babylonian and Indian fish- 
symbolism, the fish-avatar of Visnu, etc. The latter was originally a 
form of Brahma, and probably before that of some earlier deity ; in 
the Flood legend the '* fish " (jhasa) is really a horned creature, and 
a synonym of makara, which later became the characteristic vehicle 
of Varuna, but we cannot prove an early association, though such 
would be very plausible. Fish proper, especially a pair of fish, are a 
common auspicious symbol in India (e. g., in the Jaina astamangala) , 
and in at least one instance we find fish associated with a goddess of 
abundance on an early terracotta plaque.'' But on the whole, these 
analogies should not be pressed too far. 

The same applies to the dance. Folk dances, in part fertility rites 
and rain spells (e. g., the pot-dance of maidens at the Mahavrata, and 
in the Marjaliya ceremony after the horse sacrifice), and also sword 
dances (especially in Kulu) are certainly known in India, and present 
suggestive parallels with various vegetation and sword dances of 
Europe. Von Schroeder and Hertel have sought to interpret many 
Vedic hymns as being the words of early vegetation dramas or mys- 
teries, and the general Indian tendency to think of the dance as a 

^ For Dhanvantari see Gray, L. H., The Indian god Dhanvantari, J. A. O. S., 
vol. 42, 1922. 

^ See Coomaraswamy, Archaic Indian terracottas, Ipek, 1918, fig. 24. Cf. 
Dolger, F. J., IXdYC, der heilige Fisch in den antiken Religionen und im 
Christentiim ; and Pischel, Der Urspriing des christlichen Fischsymbols, Sitzber. 
Berlin Akad. Wiss., XXV, 1905. 


symbol or instigation of cosmogenic activities cannot be ignored;* 
but here, too, the available evidence has been overstrained, and the 
parallels are rather suggestive than cogent from the Grail point of 
view. The fundamental theme of the maimed king, and consequent 
cosmic disaster, necessitating a ritual designed to secure the freeing 
of the waters, is, on the contrary, plainly traceable in India, where it 
must once have existed in a more definite and unified form than that 
in which it now survives. 


Detailed studies have been made of the inakara as a decorative 
motif in Indian and Indonesian art ; ^ but little has been said of its 
significance. It may be remarked at once, that as a great Leviathan 
moving in the waters, the iiiakara is obviously a symbol of the waters 
and, as will appear from its associations, more specifically of the 
Essence in the Waters, the principle of life. The type is well known 
as the vehicle of Varuiia and the banner of Kamadeva, and it is sig- 
nificant that these deities are sometimes identified ; and as the vehicle 
of various Yaksas and Yaksis, and of the river-goddess Gaiiga. It 
occurs in the spandrils of early tympanums,^ on the architraves of 
early toranas,* and in an analogous position on throne-backs ; as a 
headdress ornament,*^ earring, or otherwise in jewellery ; isolated on 
medallions of railing pillars or cross-bars ; and very appropriately as 
a soma-sutra or gargoyle carrying away the ofi^ering-water from a 
lingam shrine." 

^Von Schroeder, L,, Gottertan^ und Weltenstehung, Wiener Zeitschrift fiir 
die Kunde des Morgenlandes, XXII, 223 ff., and XXIII, 270 ff. ; Mysterhim und 
Mimus im Rigveda, 1908; Hertel, J., in Wiener Zeitschr., XVIII, 59 ff., 137 ff-, 
XXIII, 273 ff., XXIV, 117 ft*. Cf. Coomaraswamy, The dance in India, Ency- 
clopedia Britannica, 14th ed., with bibliography, and 7 he dance of Sizfa. 

^ Cousens, H., The makara in Hindu ornament, A. S. I., A. R., 1903-04; 
Vogel, J. Ph., De makara en de voor-indische beeldhouzvkunst, Nederlandsch 
Indie Oud en Nieuw, VIII, 1924, pp. 263-276; Stutterheim, W. F., The mean- 
ing of the Kala-Makara ornament, Ind. Art and letters, NS., Ill, pp. 27-52; 
[Gangoly, O. C.]. A note on Klrtimiikha, Rupam, I, 1920; Vogel, J. Ph., 
Le makara dans la scidpture de ITnde, Rev. des Arts Asiatiques, VI, 1930. 

^ Lomas Rsi cave, best reproduction, Jackson, V. H., Not,es on the Barahar 
hills, J. B. O. R. S., XII, 1926; Vogel, loc. cit., Afb. 2; H. I. I. A., fig. 28. Later 
examples, Bachhofer, Early Indian scidpture, pis. 102, 103.; M. F. A. Bulletin, 
No. 150; Vogel, La sculpture de Mathura, pis. LV-LVII. 

* Cunningham, A., Stupa of Bharhut, 1879, pi. ix; Vogel, loc. cit., Afb. 3, 4, 5. 

''.Brandes, Le makara comme ornament de coiffure. Rev. de la Societe batav- 
ienne, 1906. 

'E.g., Vogel, loc. cit., 1924, Afb. 11; 1930, pi. XXXVIII. 


It appears in the Plant Style as the source of lotus vegetation 
(alternatively with other symbols, particularly the full vase, and 
Yaksa's mouth or navel) ; and from this type there have developed a 
great variety of decorative architectural motifs, concluding with that 
of the familiar makara-torana and the tiriivasi of a Nataraja. Not 
infrequently there are associated with the nmkara one or more dwarf 
Yaksas riding or otherwise controlling the monster; or apparently 
dragging vegetation, or sometimes an unseen object from its mouth; 
in one case it is evidently Sri-Laksml who is dragging the lotus rhi- 
zome from a makara's jaws (pi. 12, fig. 4). Sometimes a man 
or some animal or animals are represented as emerging from or 
springing from the open jaws,^ more often such forms are enclosed 
by the scrolls of the lotus vegetation which rises from the open jaws. 

The makara is always represented, at least in the early art, as a 
creature with a head like a crocodile, but with horns or fleshy feelers 
extending backwards from the end of the long snout ; with sharply 
pointed teeth ; at first with four, later with two or four rather leonine 
or dog-like legs ; and a scaly body and tail at first crocodilian, later 
ending like a fish's. Only the Lomas Rsi example shows a more 
pointed reptilian tail, with spines ; the unique example of pi. 16, fig. i, 
has dorsal spines and no legs. The later mediaeval forms, such as 
those with floriated tails in Gupta and later art, or those resembling 
land animals in Hoy.sala and later southern art, do not concern us ; 
as remarked by Stutterheim " In all these motifs the permanent ele- 
ment would be the symbolical meaning, and the mutable element the 
external form." Moreover, the original types persist to modern times 
(cf. pi. 16, fig. 4). The various forms of makara are so fully illus- 
trated that it is unnecessary to make detailed reference to the accom- 
panying plates. 

The full-face makara, which appears as an architectural motif only 
in the late Gupta period (Sarnath, Bhumara, etc.), — though earlier 
as a small metal ornament found at Taxila ' and as part of a head- 
dress at Amaravatl,^ — and finally with the designation makara-vaktra 
takes its place as the crowning element of the " caitya-window " arch 

^ E. g., Cunningham, A. S. Rep., X, p. 20, referring to a makara-torana at the 
temple of Laksmlnath at Khajuraho, " From the springing of each stalk just 
after it leaves the crocodile's mouth, a female figure hangs hy the arm, with the 
feet still in the mouth of the crocodile, as if she had sprung forth along with the 
stalk." See also pis. 4, fig. 2, and 43, fig. 6. 

■^ A. S. I., A. R., 1919-20, pi. X, fig. 31 ; cf. A. S. I., A. R., 1915-16, p. 6. 

* Bachhofer, Early Indian sculpture, pi. 129, left. 


as a fully developed makara-torana'' is also known as a klrtimukha 
(" glory head "), simha-mukha (" lion's head "), in Ceylon as kibihi, 
and in Java as kdla (-makara) and vanaspati. It is apparently an 
analogue of the Chinese tao tieh, and of Scythian, Animal Style affin- 
ity. In all probability the kirtimukha had no original connection with 
the makara, but that it was thus interpreted after the Gupta period is 
clearly established both by the term makara vaktra, and by the fact 
that vegetation or pearl garlands are regularly shown {e. g., pis. 2>7y 
fig. I and 39, fig. 2, M. F, A. Bulletin No. 167, p. 5) as hanging from- 
the jaws; as a rule the under-jaw is absent, or concealed by this 
vegetation. The same forms, for which one can hardly doubt an 
ultimately Indian origin, occur frequently in medieval European art 
(pi. 47, figs. 2,3). 

What the precise zoological prototypes of the makara are can hardly 
be positively stated. The general type is obviously crocodilian ; Vogel 
has plausibly suggested as main prototype, Crocodihis porosiis, the 
larger of the two Indian species, inhal)iting both the estuaries of 
rivers and the ocean itself ; cf . makarCdaya, the ocean. As K. de B. 
Codrington remarks, the makara " is undoubtedly indigenous." " 

I see nothing to justify Stutterheim's suggestion that the makara 
has a Hellenistic source, beyond the fact that the makara in com- 
parative mythology may be called the analogue of the Greek dolphin ; 
and this only means that each is selected as a syml^ol in its place as 
being the most obvious representative of the waters or the ocean, and 
as the king of " fish." For the Indian makara, India furnishes a 
zoological prototype; Varuna and a series of functionally related dei- 
ties provide the ultimate necessity for a '' fish," or rather " Levia- 
than " symbol; the almost certain identity of the horned jhasa of the 
flood legend with the horned makara provides literary evidence ante- 
dating the Hellenistic period, not to mention the occurrence of the 
word makara in the VajasaneyaSamhitd, XXTV, 35 ; besides this, the 
motif is only of the rarest occurrence in the art of Gandhara. For 
the kirtimukha, however, a Hellenistic origin is possible,^ though a 
" Scythian " source is more likely. 

From Bhartrhari's Nttisataka, 4, it is clear that there existed some 
legendary connection of makaras with pearls (cf . the makara as one of 
Kubera's nine treasures),* and that to extract a pearl from a makara' s 

^ Jouveau-Dubreuil, Archeologie du Snd de I'Inde, fig. 27. 

^ Codrington and Smith, History of fine art in India and Ceylon, p. 2)2>, note i. 

^ Cf. Le Coq, Bilder atlas, pp. 94, 95. 

* Kubera's " nine treasures " are the Padma, Mahapadma, Sahkha, Makara, 
Kacchapa, Mukunda, Nanda, Nila, and Kharva, and nearly all of these are 


jaws was a proverbial example of courage. In jewellery, the motif of 
pearls extended between the open jaws of niakaras is exemplified in 
Mathura sculptures illustrated by Vogel, La sculpture de Maihura, 
plates XXXIII and XXXIV, a ; later, the motif of pearl garlands sus- 
pended from klrtiinitkJia jaws is a common architectural ornament. 
In several representations a dwarf Yaksa is apparently removing 
some small object from the monster's jaws (pi. 43, fig. 7; pi. 50), 
in others an armed warrior is fighting within the jaws (Vogel, loc. 
cit., ig24, Afb. 5, 1930, p. 140, and La sculpture de Mathura, pi. X) 
and it may be assumed that the object sought is a pearl. In the 
Brhatsamhita, Ch. LXXXI, pearls are similarly said to be produced 
by the timi, another sea monster, and one of " Varuna's creatures." 
Heine-Geldern has made the plausible suggestion that pearls of such 
supposed origin may have been regarded as efficacious in the prepara- 
tion of aphrodisiacs.^ 

It should not be overlooked that the niakara, itself perhaps to be 
regarded as amphibian, is not an isolated type, but belongs to a con- 
siderable group of mythical creatures, for the most part terrestrial 
as to the head and shoulders, riverine or marine in the body and tail, 
which last is always of the same piscine form as the makaras. Two 
of these, the water-horse (jala-turaga) and water-elephant (jala- 
hastin, jalakarl, jalebha) occur often as Yaksa or Yaksl vehicles, 
particularly in the early reliefs from Jaggayyapeta (pis. 42, fig. 2, 
and 43, fig. 4). 

Despite an unnecessary confusion that has been made, the jalebha 
and makara are distinct forms and easily distinguishable.^ The best 
example of an isolated jalebha known to me is reproduced by V. 
Smith, Jain stupa of Mathura, plates LXXIII, figure i ; other unmis- 
takable examples are reproduced here on plates 40, figure 4, and 43, 
figure 4. Besides occurring thus as a Yaksa vehicle, and as an isolated 
motif, the jalebha is found also very commonly in the Plant Style as 
the source of the vegetative motif, usually a lotus rhizome (pi. 37, 
fig. 4), occasionally a series of palmettes (pi. 40, fig. 4), and is then 
clearly, like the alternative makara, Yaksa mouth or navel, or full vase, 
a symbol of the waters as the source of life. 

In addition to the water-horse and water-elephant, which are found 
in association with Yaksas or with vegetation in the Plant Style, a 
number of other creatures of the same makara-hoditd type are found 

^ Altjavanische Bronzen, 1925, pp. 24 ff. 

^As pointed out by Vogel, De makara en de voor-indische heeldhouzvkunst, 
p. 270. The jalebha is a marine monster in Brhatsamhita, XII, 13. For jalakari 
see Jacobi, Ausgew'dhlte Erzdhlungen, p. 43. 


only as separate forms, for the most part but not exclusively on medal- 
lions of railing pillars. These include the merman type (male and 
female) ; ^ the water-bull (pi. 43, fig. i) ; the water-grififon ; " lion (pi. 
9, fig. 3) ; the winged water-lion;' and probably others which I may 
have overlooked." These beings are always provided with small fins, 
exactly like those of makaras. Inasmuch as we have represented in the 
art winged, normal, and fish-tailed forms of all the chief types of ani- 
mals, that is to say, in addition to the human type, the horse,** elephant, 
lion and bull, as well as some others, it seems reasonable to assume 
that the animal species were conceived as existing in all three worlds, 
atmosphere or heaven, earth, and waters or underworld, in forms 
adapted to each habitat; just as Agni, for example, is conceived as 
existing in three forms, an atmospheric, terrestrial, and underworld." 
In the literature, at any rate, this conception of the three worlds is 
clearly the oldest, and it is also in the earliest art, surviving from an 
unknown past, that we find the animal species characteristically de- 
veloped in the three forms, winged, normal, and provided with fins 
and fish tails. 

A few words may be said about the graha, another of Varuna's 
creatures according to the Epic list. The word means " seizer." It is 
found in the Suparnddhyaya, 14, 2, synonymous with mahan hhutam 
(" monster ") of ih., 13, i, where the graha inhabits a lake to which 

* Mathura architrave, Vogel, Mathura school of sculpture, A. S. I., A. R., 
1909-10, pi. XXVII ; Bachhofer, Early Indian sculpture, pi. 104. Cf. the Baby- 
lonian " man-fish," " in Assyrian art a guardian of the tree of life and the source 
of gushing streams" (Ward, Scat cylinders . . . . , pp. 384, 410; Perrot and 
Chipiez, Chaldea and Assyria, 11, figs. 110-112). 

^ Smith, Jain stupa of Mathura, pi. LVII. 
' Smith, ib., pi. LXXV, 2. 

* The mo^ara-bodied goat, or " goat-fish " form of the Indian and Persian 
Capricornus Zodiacal sign has been discussed by Stutterheim, loc. cit., p. 36 f. 
The combination goat and fish suggests Agni, who has goat forms and is "born 
of the waters." Cf. the Sumerian goat-fish, symbol of Ea, god of the waters 
(Ward, Seal cylinders, pp. 384, 385, 399). 

® The legend that all horses were originally winged and sky-faring, but that 
their wings were cut off by Salihotra at the request of Indra is told in the Asva 
cikitsita, i, 8 (Jayadatta Suni, The Asva-vaidyaka . . . . , Bib. Ind., Calcutta, 
1887). The case is perhaps parallel to that of the dig-gajas; at any rate it would 
seem from the Karandavyuha that the winged horse (deva-)valaha originally 
meant a cloud. Cf. The great king of glory's " horse treasure " Valahaka 
(= thunder-cloud), Digha Nikaya, II, 174, 175: and Goloubew, V., Le cheval 
Balaha, B. E. F. E. O., XXVII. 

"Hopkins, Epic mythology, pp. 102, 103; Bosch, loc. cit. 


an elephant {liasti) has also repaired; ib., 14, i, is obscure, but it is 
clear from the whole passage that Garuda carries off both the elephant 
and the monster which is attacking it, with a view to devouring them. 
I cannot see any reason for interpreting either creature as itself a 
water-elephant (jalchha), as suggested by Charpentier, Supaniasage, 
pp. 234-7.'' The story in one form or another has persisted up to mod- 
ern times; in the final version (Gajendramoksa, in the Bhakta-nmld) 
A'isnu, riding on Garuda, appears to rescue the elephant, his devotee, 
from the clutches of the graha, which represents the " snare of the 
world." There is a corresponding series of representations in art, 
from which it only appears that the exact nature of the monster has 
never been clearly established ; "" thus in the Naga Jiltaka medallion 
at Bharhut (Cunningham, Stupa of Bharhut, pi. XXV) a giant crab, 
as in the Jataka text; at Deogarh (Burgess, Ancient Monuments, 
pi. 252) plainly a crocodile, evidently following the version of the 
Bliagavata Puraija; at Koramangala (Mysore A. S. R., 1919-20, p. 5, 
and pi. Ill) likewise a crocodile (very like a uiakara) ; so also in the 
case of the original which must have formed the basis of the cover 
design of Kipling's Just So Stories; and in later paintings {e.g., 
Rajput painting, 1916, pi. XVI) usually a " laithly worm " or hydra. 
The problem is not of immediate importance from the present point 
of view. 

Varuna : the word niakara occurs in Vedic literature only in the 
Vujaseyam Sainhitd, but the later synonymy of jhasa-ketana with 
niakara-kctana ( = Kamadeva) suggests that the horned jliasa of the 
Flood legend (SBr., I, 8, i) and the jJiasa of ib., XIII, 6, 2, 20, may 
be equated with the inakara, though both occur in the Epic in the list 
of sea-monsters (//////, inakara, jhasa, kiirma, grdha, etc.) that are 
" Varuna's creatures." ^ 

^Unless because the jalakarl in the story of Nami (Jacobi, Ausgcwdhlte 
Erzahlungcn, p. 43) has the character of a " seizcr." 

^ The Epic versions of the story make the graha a tortoise. Monier-Williams, 
Sanslzrit Dictionary, s. v. graha, outdoes all Indian sources in the variety of his 

On the water-elephant, see also Zimmer, H., Spiel nm dm Elcfanten, Berlin, 
1929. P- 38, note 2. 

^ Sinhimara and sisumara, dolphin or porpoise, occurs in RV., AV., etc. (with 
jasa (sic) in AV., XI, 2, 25) ; and in the Taittiriya SamJiita, where it is glossed 
as graha by Sayana. Here -mara perhaps — mateara, cf . Kusinara — Kusinagara. 


Later, Varuna is regularly makara-vdhana'^ {e.g., Agni Piirdna, 
Ch. LI), though makara-ketu in the Visnudharnwttara, III, 52. It is 
not implausible to suppose that a sea-monster — jhasa or makara — 
was a symbol or form of Varuna at a period earlier than that for 
which we have positive evidence ; the later vdhanain may have been 
originally a theriomorphic form of the deity in person. 

It is well known that the fish avatar of Visnu in the Puranas de- 
rives from the " fish " form of Brahma-Prajapati (SBr., I, 8, i, and 
in the Epic) ; but in the Brfdimana account, " fish " is a misnomer, 
for the jhasa is horned, and " sea-monster " would be a better ren- 
dering. That the myth may well be one of much greater antiquity is 
shown by an allusion to the flood legend in AV., XIX, 39, S."" Hence 
it is far from implausible to suppose that the sea monster (jhasa or 
makara) was originally a form or symbol of Varuna. The case of the 
tortoise is analogous to that of the " fish ": the form is assumed by 
Visnu to support Mt. Mandara at the Churning of the Ocean ; in 
SBr., VII, 4, 3, 5, and VII, 5, i, 5, Prajapati assumes the form of a 
tortoise, moving in the waters, to accomplish the work of creation 
(the Churning of the Ocean is of course a creation myth, and we have 
seen reason to suppose that it is referred to already in the Rg Veda) ; 
in the Satapatha BrdJimana, further, the tortoise laid down on the 
fire-altar, between layers of avakd plants representing water, is iden- 
tified with rasa, with the three worlds (sky, air, earth), with Praja- 
pati, with the sun, and with the breath of life ; in the Vdjasaiieya 
Saihhitd, XIII, 31, the tortoise is called "lord of waters"; in AV., 
XIX, 53, 10, as Kas3'apa, the tortoise is again identified with Prajapati 
and called " self-existent " ; a previous identification of the tortoise 
with Varuna may be reasonably inferred.'' 

^ Badami reliefs, 6th century, Banerji, R. D., Bas-reliefs of Badanii, Mem. 
A. S. I., 25, pi. XI, e; XXI, c; XXIII, a and e; also p. 34 above and pi. 45, 
fig. I. Masrur, Kangra, 8th century (with ram's horns), A. S. I., A. R., 
1915-16, pi. XXXIV, a; also Bhattacharya, B. C, Indian images, pi. XVIII. 
Lingaraja temple, Bhuvanesvara, 13th century, A. S. I., A. R., 1923-24, pi. XL. 
As blja of the Tantrik Svadisthana Cakra below the navel : " within it is the 
white, shining, watery region of Varuna .... seated on a makara," Avalon, 
Serpent power, p. 138 and pi. Ill, and Pt. II, p. 38 f.). 

^The interpretation is, however, questionable, see Whitney, Atharva Veda, 
H. O. S., vol. 8, p. 961. For the flood legend generally see Hohenberger, A., 
Die indische Flutsage und das Malsyapurana, 1930, and Winternitz, M., Die 
Flutsagcn des Altertums iind der Naturvolker in Mit. Anthrop. Ges., Wien, 
XXXI, 1901. 

^ Cf. also Yajnavalkhya, I, 271 ff. and JBr., Ill, 272 (Akupara, cosmic tortoise). 
While the " fish " survives as the vehicle of Varuna, the tortoise becomes that 
of his river-consort Yamuna. 


Similarly in the case of the boar, the third avatar of Visnu, who 
assumed this form for the purpose of raising the earth from the 
cosmic waters at the commencement of the Varaha Kalpa. In the 
Linga Pur ana the tradition is preserved that it was Brahma who 
slept upon the waters, determined to create, assumed the form of a 
bear, and raised up the earth ; so also in the Ramdyana. 

In SBr., XIV, i, 2, 11, it is Prajapati who assumes the form of a 
boar (cniusa) and raises the earth from the cosmic waters; the earth 
is called his *' mate and heart's delight," and we have clearly to do 
with the cosmic deity and Mother Earth, easily recognizable as cor- 
responding to Varuiia and Aditi. 

Kamadeva: is identified in the Epic with Pradyumna, son of 
Krsna; in the Vumidharuiottara, III, 52, the same Pradyumna and 
his wife Rati are identified with Varuna and Gauri " holding a lotus 
of dalliance," and Varuna himself has a iiiakara kctii, not vdhana. 
Pradyumna-Kamadeva is likewise in the Epic luakara-dhvaja or -ketu. 
In mediaeval texts Kamadcva's constant epithet, synonym, or at- 
tribute is niakara-dJivaja {e.g., Visuudhannotiara, III, 73, 20-24), 
and this is also the name of an aphrodisiac advertised to the present 
day. It should not be forgotten that Kamadeva is a Yaksa {Utta- 
rddhyayana Tlka, Jacobi, p. 39) and identical with the Buddhist Mara 
(Buddhacarifa, XIII, 2). Kama is also a form of Agni, and Agni is 
born of the waters. 

Evidence of an earlier cult of Pradyumna-Kamadeva has been 
recognized in the Besnagar viakara-dhvaja of Suhga date (pi. 16, 
fig. 2, and pi. 45, fig. 3)^ and the example of unknown provenance 
but probably similar date here reproduced for the first time (pi. 16, 
fig. i) may have the same significance, the mortice showing that this, 
too, was a standard. Kamadeva with the makara-dhvaja is actually 
represented witli Rati at BadamT,^ at the Kailasanatha, Elura,^ and 
elsewhere ; in the Kddanibari his image is referred to as painted on a 
bed-room wall; in two Gandhara reliefs his daughters bear a uiakara 
standard," and twice at Sarnath the same holds good for one of his 
attendants in a Mara Dharsana scene.^ 

^ Cunningham, A. S., Rep., X, p. 42 and pi. XIV; Chanda, R., Archeology and 
Vaisnava tradition, Mem. A. S. I., 5, 1920. The Besnagar makara may have had 
a rider, as suggested by Bhandarkar in A. S. I., A. R., 1913-14, p. 191, but if so 
it must have been of the dwarf Yaksa type, common in reliefs, and certainly not 
a Garuda. 

^Banerji, R. D., Bas-reliefs of Badami, Mem. A. S. I., 25, 1928, p. 34. 

'Burgess, A. S. W. I., V, 1883, pi. XXVI, 2. 

* Foucher, A., L'Art greco-bouddhique dii Gandhara, II, p. 196, figs. 400, 401. 

^ Sahni and Vogel, Catalogue . . . . , C(a)i and C(a)5, pp. 184, 191. 


Inasmuch as the makara generally means the waters, hence more 
specifically the essence in the waters (rasa in its various equivalents, 
sap, semen, Water of Life, etc.), and virility (vtrya, Visnudharmot- 
tara, III, 52), the association of the makara with Kamadeva or any 
deity of fertility is quite appropriate. 

It has already been remarked that the epithet jhasa-ketana both in 
Sanskrit and Hindu sometimes replaces the usual makara-dhvaja; 
Monier Williams cites Kuvalayananda, 33, with the double meaning 
" god of love " and '' the sea." Hence it is perhaps significant of a 
quite early association of the sea monster with Pradyumna-Kamadeva 
that in the list of symbolical victims, SBr., XIII, 6, 2, 20, there is 
assigned to the jhasa a " sportive woman." 

Yaksas and YaksTs : nearly all the vehicles by which these vegetal 
divinities, the primary theme of the present treatise, are supported, 
can be directly or indirectly shown to be connected with the waters, 
and this forms part of the evidence available for the view (see p. 34 
above) that Yaksas and YaksTs should be identified with the Gandhar- 
vas and Apsarases as originally conceived, that is to say as primarily 
connected with the waters, and secondarily with vegetation. 

Even the horse (Yaksas, I, pi. 5, fig. i), it will be remembered, is 
"water-born" and connected with Varuna: and the elephant (ib., 
pi. 3, fig. 2, and pi. 4, fig. 2) may be a sky elephant, that is to say, a 
cloud. On the other hand, the dwarf Yaksa vehicle (ib., pi. 3, fig. i, 
and pi. 4, fig. i) seems to represent a gnome or earth spirit, cf. the 
Yaksa Atlantes, here plate 8, figure 2. But it is noteworthy that the 
makara is the commonest of all the vehicles (ib., pi. 6, figs, i and 2, 
pi. 19, and see p. 47 above) and next in frequency are the fish or 
makara-tailed forms of terrestrial animals, particularly the water- 
horse (jala-turaga) and water-elephant (jalebha, jala-hastin) ; the 
latter form occurring also in the Plant style only less often than the 
makara as the source of lotus vegetation. 

As noted below, when the nadi-devatds, who in the earliest repre- 
sentations are primarily distinguished by carrying a punna-ghata 
(never a Yaksa attribute) are later specifically differentiated as 
Ganga and Yamuna, the former retains the makara, the latter is given 
a tortoise. Iconographically the differentiated forms of the river god- 
desses (in northern India only) is directly derived from that of the 
Yaksi-dryad, and this implies that the latter, despite the vegetal and 
apparently terrestrial habitat, was still primarily a spirit of the waters. 


Other deities : In at least one case the Moon, in a Navagraha 
group/ has a luakara vehicle, and this is comprehensible through the 
common identification of the mbon with Soma and the close connec- 
tion of both with Varuna. Parvati, at Elura, as Uma performing the 
pahcagui fa pas, is represented with a makara vehicle." There is a 
curious Pala figure, perhaps also Parvati, at Jamir, Monghyr District, 
Bengal ; ^ the goddess is seated on a lotus, has a lion cognizance, and 
is four-armed, with a cup in the lower right hand, /r/i/l/a-handled bell 
in the upper right, a niakara-dJwaja in the upper left, and a nude 
child in her lap supported by the lower left. There is also a rare coin 
of Samudragupta, with a standing goddess on a makara vahanam, 
with a long-stalked lotus in her right hand, and a crescent-topped 
standard beside her ; Burgess called her Parvati, Allan says Ganga,* 
and either identification is possible. That Parvati is sometimes called 
the sister of Ganga, that in the Agiii Pumna, Ch. LXIV, she is said 
to accompany Varuna, and VaruiiT is sometimes replaced by Gauri, 
offers perhaps sufficient explanation. According to Monier Williams, 
Varuna vi is a synonym of Laksmi. The goddess with a makara 
vehicle published by Vogel ' may be Parvati or Gauri, or possibly 
Varuiii. The only other deity who to my knowledge is connected with 
the makara is the ninth Jaina Tu'thankara, who has this cognizance. 


It is not intended here to present an exhaustive account of the 
place of the lotus in Indian culture and art, but only to discuss the 
points that are most relevant to the present enquiries. Texts already 
cited (above, p. 23) from the Safapallia Brahmaua show that the 
lotus was primarily understood to represent the Waters ; secondarily 
also, inasmuch as the flower and still more obviously the leaf rest on 
the waters, the earth — for the earth is conceived of as resting on the 
back of the waters, and supported by the waters, which extend on 

^I have the photo, but have mislaid the reference. In another Navagraha 
group from Bengal, published in the A. R. Varendra Res. Soc, 1928-9, the 
Moon's vehicle is also probably a makara, but the animals are not easily recog- 
nizable. Varuna is regarded as the presiding deity of the Moon (see Bhat- 
tacharya, B. C, Indian images, p. 32). 

' Burgess, A. S. W. I., V, pi. XXX, 2. 

^ Burgess, Ancient monuments, pi. 225. 

* Burgess, A. S. W. I., II, pi. VII, 2; Allan, Cat. coins Gupta dynasties, Brit. 
Mus., pi. II, 14 (a true makara, not elephant-headed as stated ib., p. 17). 
Hopkins, Epic mythology, p. 118; also Visnudharmottara, III, 52. 

^ De makara en de voor-indische heeldhouwkunst, p. 273. 


either side of it. These related and by no means far-fetched inter- 
pretations sufficiently account for the use of the expanded lotus flower 
in iconography and architecture as the typical basis or support of a 
figure or building. I allude here ( i ) to the familiar padmasana and 
padma-pltha of Indian images and the corresponding " bell-capitals " * 
of supporting columns, and (2) to the usual lotus petal mouldings of 
architectural basements, whereby it seems to be implied that the whole 
building is supported by a widely extended lotus flower, that is to say, 
by the earth, and in the last analysis by the Waters. Furthermore, 
the lotus is represented as a direct source of wealth, as in the case 
of the padma nidhi of Kubera (pi. i, and pi. 46, fig. i ) , and the ratana 
manjankds of Bharhut and Saiici (p. 4 and pis. 11 and 13). These 
meanings and values do not at all exclude that of the implication of 
birth in the Waters, conspicuous in the case of Srl-Laksmi, who is 
the earliest divinity to be constantly represented with padma-pltha or 
padmasana, though in the case of other deities not so directly born 
from the waters, the idea of support seems to be indicated rather 
than that of *' divine birth," which has hitherto been the usual inter- 
pretation ; on the other hand, the more edifying symbolism of purity, 
drawn from the fact that the lotus leaf is not wetted by the water 
that it rests on, nor is the flower soiled by the mud from which it 
springs, belong to a later cycle of ideas, and only come in with the 
sectarian, Buddhist and devotional developments. 

Our attention is next called to the fact that in the early " decora- 
tive " art, which from our point of view should rather be regarded 
as an iconography of the Water Cosmology, the Plants, whose virility 
and healing powers are so much stressed in the literature, are almost 
invariably represented by the lotus, no doubt because of its directly 
evident origin in the Waters. So, too, the lotus represents the Tree 
of Life; this cosmic tree which sprang originally from the navel of 
Varuna, bearing the deities within its branches (presumably thought 
of as those of an actual tree), when later it is represented (in the 
Mahdbhdrata and in late Gupta and early medieval art, see above, 
pp. 2, 3) as rising from the navel of Narayana or Visnu and bearing 
Brahma (pis. 11, fig. 4, and 47, fig. i), has always the form of a 
lotus, whence Brahma's epithets Abjaja, and Abjayoni, " born of the 
water-born," i. e., of the lotus. 

Except in the case of the lotus medallions, representing the upper 
surface of a single flower, it is the whole lotus plant that is generally 

*I have dealt with this point fully in Early Indian iconography, in Eastern 
Art, I, p. 170, and in the Indian Historical Quarterly, VI, 2>72)' 


represented in art. This whole plant, in Nature, consists of a rhizome, 
with nodes at regular intervals, each node provided with small scale 
leaves and rootlets, and giving rise to numerous larger leaves and 
flowers which rise to the surface of the water ; in other words, there 
is a creeping submerged root-like stem which throws off flowers and 
leaves at intervals, but there is no branching stem, and the stalk of 
each flower or leaf rises directly from the rhizome. Bearing these 
facts in mind, it is easy to recognize in the ordinary lotus spray, 
whether rising from a vase of plenty and/or forming a vegetative 
meander springing from a vase, a conch, a inakaras jaws, a Yaksa's 
mouth, or a Yaksa's or Visnu's navel, a portion of the whole plant ; in- 
numerable examples of all these types are illustrated in the accompany- 
ing plates. As to these points of origin, we have seen that the navel 
is regarded typically the procreative center,' and all the rest imply and 
represent the Waters. The majority of these characteristic points of 
origin persist in the art from the earliest to modern times (cf. pi. 5, 
fig. 2) ; at the same time the lotus prototype can be recognized even 
when decorative modifications of the vegetative forms result in motifs 
no longer obviously of lotus origin, for the nodes are always clearly 
traceable. An extensive work on Indian decorative art is much 
needed ; and so far as the vegetative ornament is concerned, such a 
book would be almost entirely occupied with forms of obviously or 
derivately lotus origin. The palmettes, for example, so characteristic 
at Bharhut and Sanci, consist of lotus leaves and flowers rising from 
a single node (pis. 43, fig. 5 ; 44, fig. 4) ; the purely Indian acanthus- 
like motifs of Andhra and Gupta art are directly derivable from 
simpler forms of ribbed and folded lotus leaves seen at Sanci ; and 
finally, the remarkable garland motif so magnificently developed at 
AmaravatT is nothing- more than a decorated lotus rhizome. 

The last formula, that of a '* garland " borne by dwarf or normal 
Yaksas, or more rarely by Yaksa mithunas, needs to be considered at 
somewhat greater length. The true nature of the motif in its later 
forms is not immediately evident; Vincent Smith called it a bulky 
tinsel roll. By him and others it has been regarded as a form of the 
Roman and Syrian, especially Alexandrian, motif of a garland borne 
by Erotes, though as remarked by Vogel, " by what route or means 
this popular motif reached India is still a mystery."" Vogel at the 

^ The other observed sources from which there spring lotus meanders are 
(i) the water elephant (jalebha), (2) what is apparently a terrestrial ele- 
phant, but as shown above, p. 4, is more probably intended for a sky elephant 
or cloud. 

^ Smith, V. A., History of Fine Art in India and Ceylon, ed. I, p. 384; Vogel, 
J. Ph., La sculpture de Mathura, pp. 79-81. 


same time mentions only incidentally one of the main clues to its real 
character, calling attention to the fact that the " bizarre garland," 
typically at Amaravati, but also at Mathura (I, 3 in the Mathura 
Museum), is frequently made to issue from the open jaws of a 
makara. To be more precise it issues or is dragged forth by Yaksas,* 
or in one case by Sri-Laksmi, from the open jaws of a makara (?. e., 
from the waters), or alternatively, from the open mouth of a dwarf 
Yaksa, implying an origin from the mukhya-prdna, or from the wa- 
ters, as observed above, p. 24. Now these curious and certainly not 
accidental points of origin of the " garland " are precisely two of the 
four common points of origin of the normal and unmistakable lotus 
meander, as met with at Bharhut, Saiici, at Amaravati, and later ; we 
have remarked already that the lotus meander is a rhizome, with 
leaves and flowers springing from the regularly spaced nodes, as in 
Nature. And further, our " garland " itself is provided with nodes 
at regular intervals, and though at Amaravati realism is neglected in 
the interests of decorative symmetry, the scale leaves being dupli- 
cated, so that the indication of direction of growth is lost, a com- 
parison of the forms presented by the node and scale leaf successively 
at Bharhut, Saiici, and Amaravati (pi. 40, figs. 1-3, cf. pi. 37, figs. 3 
and 5) presents us with an altogether convincing evolutionary series. 
It is thus beyond doubt that the motif in question is really that of 
a lotvis rhizome originating in the waters and borne by Yaksas." But 
whereas at Bharhut and Safici the rhizome itself is represented simply 
as a smooth round " stem," the stem at Amaravati is most elaborately 
decorated, and the representation is modified in the interests of sym- 
metry and of the spaces to be filled with figure groups, with omission 
of the flowers and leaves, duplication of the nodal scale leaves (pi. 40, 
fig. 3) ; and in some cases the termination of the garland is made to 
correspond to its origin, so that it seems to enter at one end the mouth 
of a dwarf Yaksa or the jaws of a makara. The symmetrical arrange- 
ment last referred to is often to be seen at Amaravati in the case 
of the normal lotus meander (pi. 38). It may also be remarked 
that while the lower nodes of the roll are more formally ornamented 

^Vogel, Le makara dans la sculpture de I'Inde, p. 141, ascribes all these fea- 
tures to the " fantaisie du sculpteur indien." This seems to me contradictory to 
the whole character of Indian art — the more we know about it, the more its 
formulae reveal, not indeed " quelque sens mystique," but certainly " quelque 
sens symbolique," i. e., definite in significance. Cf. above, p. 13, note i. 

*This was first pointed out in Early Indian iconography, II, Eastern Art, I, 
pp. 187/8. 


and less realistic than the upper, one naturalistic feature, that of the 
rootlets, is markedly developed (pi. 37, fig. 5). 

The symmetrical tendencies above alluded to result sometimes in 
a treatment of the lower node such that it is made to consist of paired 
addorsed approximated inakara jaws, the bodies being omitted, and 
the " garland " proceeding from the open jaws both to right and left 
(pis. 4, fig. I, and 37, fig. 5) ; it would be more correct to say that in 
such cases the node is replaced by two points of origin, and that the 
direction of growth is ignored. It is conceivable that by this time a 
consciousness of the significance of the motif had been lost, as we may 
assume to have been the case later when the formula of addorsed 
makara jaws occurs in Gupta and medieval art (pi. 39, fig. 2). 

The Mathura examples (Smith, loc. cit., pi. LXXXVII, C; Vogel, 
loc. cit., pis. V, a, and LX, a ; here pi. 12, fig. 3) provide us with inter- 
mediate and less elaborately decorated or symmetrically modified forms 
than those of Amaravati, and occupy their natural place in the chrono- 
logical stylistic development, offering at the same time additional 
evidence of the close connection between the art of Amaravati and 
that of Mathura during the middle and latter part of the second cen- 
tury A. D., estal)lished on other grounds by Bachhofer." 

The formula occurs also in Gandharan art, generally as an imbri- 
cated roll-garland borne by Erotes, with a total omission of the 
nodes ; the Erotes usually face each other in pairs, instead of moving 
in one direction as in all Indian examples. The Gandharan examples 
cannot be exactly dated, but there is no reason to suppose that any 
antedate the second century A. D. 

It is not proposed to discuss here the relation of the Indian and 
Western Asiatic examples of the '' garland " ; it will be seen, how- 
ever, that the motif may well be of Indian rather than of western 
origin, and in view of the other very plain traces of Indian influence 
that have been recognized in Alexandrian art, there would be nothing 
surprising in this.' 

^ Bachhofer, L., Early Indian sculpture, pp. 61, no. 

" Cf . Berstl, H., Indo-koptische Kunst, Jahrb. as. Knnst, I, 1924; Dimand, M., 
Indische Stil-Elemente in der Ornamentik der syrischen und indischen Kunst, 
O. Z., IX, pp. 201-215, and Die Ornajnentik der iiyyptischen Wolhuirkereien, 
1924; Strzygowskij Les elements proprement asiatiques dans I'art, Rev. des 
Arts Asiatiques, VI, 1930, p. 33 — " rexpansion de I'art asiatique en Europe, 
expansion a laquelle Alexandre avail si largement ouvert les portes . . . . et qui 
apporta sur la Mediterranee le patrimoine artistique du mazdai'sme." 



We shall drain the well full of water, 

That never is exhausted, never faileth. 

— RV., X, loi, 5 and YV., IV, 2, 5. 

Throughout the history of Indian art the full vessel (purna kalasa, 
punna ghata, etc.) is the commonest of all auspicious symbols, em- 
ployed equally by all sects, and occurring not only in India proper, 
but also in Farther India and Indonesia. 

The earliest examples are found at Bharhut and SaiicT, in connec- 
tion v^ith the representations of Sri-LaksmT ; w^e find (i) the goddess 
standing or seated on a lotus, (2) the same, but the lotus rises from 
a punna ghata, and (3) the pimna ghata alone, with a mass of lotus 
flowers and leaves rising from it. The three types are apparently 
equal and synonymous symbols of abundance, and it may be that the 
vase alone should be regarded as an aniconic symbol of and equivalent 
to the goddess herself.^ In Jaina art the punna ghata is one of the* 
Astamangala, or Eight Auspicious Symbols (pi. 31, fig. 2) and also 
one of the fourteen lucky dreams of Tisala. The full vessel is carried 
as a symbol by some divinities, e. g., by Nagas (pi. 33, fig. 4) and by 
river-goddesses. It is used in the worship of deities and the reception 
of human guests and in the festival decoration of cities and shrines.' 
In the case of shrines also, a pair of full vessels are commonly placed 
at entrances, as constantly seen for example, on Amaravatl reliefs 
representing stupas ; or a frieze may consist entirely of a row of full 
vessels represented in relief. As an integral architectural motif it 
occurs in rich and varied forms as an essential part, generally the 

^ For further illustrations, see my Early Indian iconography, II, Sri-Laksml, 
in Eastern Art, I, 1929, pi. XXIV. Cf. Dhisana as (i) a goddess of abundance 
and a "mother," and (2) as soma-vessel and figuratively soma-juice (Johann- 
sen, K. F., Ueher die altindische Gottin Dhisana . . . . , pp. 26-28). 

^Innumerable examples could be cited from the literature of all periods and 
sects from the Sutra period onwards ; I cite only a few ; viz, from the 
Bee-song of Surdas, when the gopis are welcoming Udho, they " set before him 
full golden jars, and circumambulated him"; in Mahdvaiiisa, XXXI, 40, "A 
thousand beautiful women from the city, with the adornment of fair full vessels 
{supunna-ghata-bhusayo) surrounded the car containing the relics"; Manimek- 
halai, Bk. i, "Do therefore decorate the city, the great royal roads, and the 
halls of faultless learning; put in their appropriate places full jars, seed-vessels 
with budding sprouts, and statues holding lamps "; AV., Ill, 12, 8, at the dedica- 
tion of a house, "Bring forward, O woman, this full jar" and XIX, 53, 3, 
where a full vessel is "set upon time." In the Harsacarita, VIII (§ 227), "a 
golden vessel adorned with sprays " is set on the altar of a Brahmanical tem- 
ple. For the use in modern Brahmanical ritual see Burgess, J., The ritual of 
Rdmesvaram, Indian antiquary, XII, 321, where a decorated kiimhha represents 
king Varuna. 


capital or sub-capital, of monolithic or structural columns (pis. 17, 
figs. 2, 3; 32, fig. i) or as the support of a pilaster (pis. 27, fig. 2; 
42, fig. 2). It constitutes the well-known pot-and-foliage capital of 
medieval Indian architecture, a form that has generally been regarded 
as a development from the old " bell " capital ; but while it is possible 
that the capital as such has originated in this way, this must not be 
thought of as an origin of the motif itself, which is already fully 
developed in Suiiga art. 

The pfinia kalasa is plainly thought of as an inexhaustible vessel, 
but the actual form, always associated with vegetation, should I think 
be clearly distinguished from that of the plain jars sometimes carried 
by the early undifferentiated river goddesses, and also from that of 
the anirta phial borne by Indra and some other deities, though these 
simpler vessels likewise are of necessity thought of as inexhaustible. 
As seen in outline or relief, the purna-kalasa is generally a globular 
vessel with a foot, and a constricted neck; the body of the vessel is in- 
variably encircled by a ribbon or other band, tied with knots and serv- 
ing the purposes of a magical '' fence " (see J. A. O. S., vol. 48, 
p. 2y2i) ; from the mouth there rises a spray or bunch of lotus flowers 
and leaves, almost invariably so arranged that a pair of flowers or 
leaves hang over symmetrically on each side of the mouth, like the 
volutes of a palmette. Very commonly, and especially when narrow 
vertical spaces are available for the reception of symbolic ornament, 
the vegetative element is extended upwards to a considerable height, 
either as a conventional candelabra-like tree, or as a long spray of 
lotus, bearing flowers and leaves, and enclosing or framing birds and 
beasts in its convolutions (for some of these types, see pis. 14, fig. 2, 
and 42, fig. I ) . 

Thus the form is essentially that of a flower vase, combining a 
never-failing source of water with an ever-living vegetation or tree 
of life. The type is of the widest distribution in later art, and it 
can always be identified by the symmetrically placed lateral over- 
falling leaves or flowers. Examples are common in the art of the 
Renaissance, and the Persian vase carpets may also be cited ; these 
forms must originate either from the Indian, or from cognate forms 
in Western Asia, if such existed equally early. 

The vase of plenty described above is clearly a life symbol, and 
the formal offering of such a vase can only be the expression of a wish 
that the recipient, or in general all those present, may enjoy health, 
wealth, and long life. The representation in art implies similarly a 
desired instigation by suggestion of all the vegetative energies in- 


volved in the current conceptions of well-being ; as a symbol it clearly 
belongs to the order of ideas characteristic of the ancient life cults 
of fertility and fruitfulness. 

As the motif has had a continuous history from the Suiiga period 
onwards, and there is not the slightest reason to suppose that it was 
either invented or borrowed precisely at the moment when stone came 
into use as a building material, we must infer an antecedent history 
of the motif on Indian soil. It is further, a general law that the 
farther we go back in tracing any Indian ornamental motif, the clearer 
become its character and meaning, and it may be said, since we are 
unable as a rule to approach a period of definite beginnings, that to go 
back still farther would lead us to still more definite and consciously 
employed forms. This is no more than a parallel to what has already 
been recognized in the literature, where from later sources we recover 
trace of a once more consistent mythology and ritual of the chthonic 
vegetative powers. 

Since we cannot expect to recover many actual documents of pre- 
Maurya art in impermanent material, particularly wood, it will be 
pertinent to call attention to the Mesopotamian analogy of the Flow- 
ing vase, which gradually developed into a vase of vegetation ; for a 
similar evolution may very well have taken place in India. In the 
representations of this '* merveilleux symbole qui etait comme le Saint- 
Graal de I'epopee chaldeenne," to quote the words of one of the 
greatest scholars of Sumerian antiquities, there can be recognized an 
" evolutionary " and more or less chronological sequence of types. 
At first there are plain globular vases, held by standing or seated 
personages, one hand below, the other on the vase (Heuzey, loc. cit., 
pi. V). Then comes the typical and very beautiful form, that of a 
vase from which spring two undulating streams of water, to right and 
left ; these are held by male or female genii of the waters, represented 
in sculpture or metal-work, e. g., the beaker of Gudea," or by a divin- 
ity represented on seal cylinders, e. g., Heuzey, loc. cit., p. 41, and 
Ward, Seal cylinders, Nos. 286, 650, etc., in some cases numbers of 
such flowing vases may be arranged symmetrically to form an all- 
over design (pi. 41, fig. 3). Occasionally a small vegetative sprout is 
shown between the two rising streams, and this later develops into 
an ear of corn. Sometimes four streams are represented ; very often 
accompanied by fish, perhaps as a symbol of Ishtar, or simply to 
emphasize the sense of the water. Finally we get a vase of a somewhat 
different shape, having a tall central sprout and two lateral volutes, 

^Unger, No. 47. 


which seem to represent the original streams of water (Ward, No. 
203 ; here pi. 27, fig. i) ; these vases are offerings set before a deity 
(Heuzey, p. 163; Unger, No. 59; Ward, Nos. 421, 1235). We thus 
arrive at a form at least analogous to the Indian, inasmuch as it is a 
vase of vegetation, with symmetrical over-falling volute-like elements 
on either side ; and it may be suggested that perhaps the Indian form 
has been developed from an older type of actually flowing vase, 
analogous to that of the early Chaldean art. 


The Mathura school of sculpture has yielded numerous examples 
of large ornamented l)owls supported by figure pedestals ; the sup- 
porting figures are cither Yaksa or Yaksi groups, or a form suggest- 
ing Sri-Laksmi or some analogous goddess of plenty. Two of these 
monolithic bowl pedestals, known as the Stacy and the Pali Khera 
groups, have often been described, and have been recently discussed 
again by Vogel,^ in relation to several other pieces, viz., an inscribed 
bowl from the Pali Khera site, a bowl-bearing head "" from the Jamna 
Bagli site, and the complete female figure now in the Museum of the 
Bharata Kalfi Parisad at Benares. A number of other examples arc 
known : ( i ) a damaged bowl supported by four female figures wear- 
ing heavy anklets and more or less inebriated, the whole probably origi- 
nally a little over two feet in height," {2) the Kota group now, like 
the last, in the Indian Museum, Calcutta; the pedestal consists of 
" two females standing side by side under a large (asoka) tree, which 
is fully represented at the back of the stone." This group was de- 
scribed by Cunningham " who further remarks that " it is quite pos- 
sible .... that the top of each of these Bacchanalian groups was 
only a hollow bowl," (3) a badly damaged group of Mathura origin 
found at Tusaran Bihar, the pedestal consisting of eight figures, 
almost or quite nude, and apparently intoxicated ; leaves falling over 
the shoulder of one of the male figures show that the group was repre- 
sented as standing beneath a tree, and Cunningham remarks that " it 

^ La sculpture de Mathura, Ars Asiatica, XV, 1930, pp. 52-56, " Porteurs de 

^ Cf. the bowl-bearing head in a Brahmanical cave at Lonad, A. S. I., N. I. S., 
V, 1883, pi. XLV. 

^ Chanda, R. P., The Mathura school of sculpture, A. S. I., A. R., 1922-23, 
p. 167 and pi. XXXVIII, b. 

' Cunningham, A. S., Reports, XX, pp. 48-50. This piece seems to have been 
sheltered by a "four-pillard mandapa." 


is almost certain that these groups may have formed the support of a 
bowl " ; ^ (4) the figure of Sri-Laksmi, B 89 in the Lucknow Museum, 
here pi. 49, may very possibly have supported a bowl.^ 

The character of the pedestal figures in all these cases would ap- 
pear to be anything but Buddhist, and is not in fact Buddhist. But 
we are already accustomed to the constant presence of un-Buddhistic 
representations in connection with Buddhist monuments, and in any 
case the inscriptions preserved on two of the pieces, viz., the Pali 
Khera bowl {samghiyanam parigaha ....,'' for the acceptance of 
the community ") and the Jamna Bagh head (dedication to the Suvan- 
nakara-vihara ) prove beyond doubt a Buddhist use. Vogel, loc. cit., 
p. 54, has suggested that the vases supported on pedestals represent 
the Buddha's begging bowl, and were set up to receive offerings of 
the faithful, and cites a similar practise in Burma. Against this we 
have the facts: (i) the vases are unlike a begging bowl (pinda- 
pdtra) in form, (2) the ornamentation of a begging bowl is explicitly 
forbidden by Vinaya rule, and (3) the associated figures on the 
pedestals are always connected with the idea of a liquid, either an 
intoxicating liquor or pure water; they are in fact genii of living 
waters, either Yaksas, Yaksis, or forms of or related to Sri-LaksmT. 
Certain of these facts suggest a comparison with a Chaldean font 
described by Heuzey ; ^ in any case, they suggest that the bowls were 
meant to contain water. 

Rejecting several other possibilities hardly compatible with Bud- 
dhist usage,* it seems to me far more plausible to suppose that we have 
to do with water bowls than with alms bowls. Such water bowls, 
called dcamana-kumbhi or dcdma-kumhhi were regularly placed at 
the entrances to Buddhist shrines, to hold water for washing the 
hands and feet of the visiting worshipper.'' This interpretation, more- 
over, better accords with the historical tradition suggested by the 
Babylonian " bassins," cf. Heuzey, loc. cit., p. 151. 

Whether interpreted as pinda patras or as dcamana-kumhhis it 
seems a little strange that the bowls or figure pedestals should be 

^ Cunningham, A. S., Reports, XI, p. 65 and pi. XX. 

"^For this figure, see Cunningham, A. S., Reports, I, p. 240 and pi. XL; my 
H. I. I. A., fig. 74; Early Itidian iconography, II, Srl-Laksnu, in Eastern Art, 
I, 1928-9, fig. 22; Vogel, La sculpture de Mathurd, pi. L. 

^ Le bassin scvdpte et le symhole du vase jailUsant, in Les origines orientales 
de I'art, pp. 149 ff. 

*E.g., the bowl (kunda) to be used as a rain gage (varsamdna) in front of 
a granary, Kautilya, Arthasastra, II, 5. 

^ Mahdvagga, I, 25, 19; Cullavagga,Y, 35, 4; Thupavamsa, LIV, 2; Geiger, 
W., Mahdvamsa (translation, p. 185, note 3). 


confined to the Mathura school and Kusana period — though the few 
known examples of stone yfipas provide a parallel case. It hardly 
seems as though such an elaborate, sophisticated and unique form 
could have been invented suddenly, or that the constancy of type of 
the pedestal figures — genii of vegetation, liquor, and abundance — could 
be accidental ; nor can the functional necessity have existed only a 
brief period. It may perhaps be inferred that earthenware water bowls 
had previously been set up at the entrance to Buddhist or other shrines 
(in all such ritual matters the Buddhist cult inherits rather than 
invents) upon carved wooden pedestals ; and that later, after the decay 
of Buddhism, a simple earthenware bowl without any elal)orate stand 
must have sufficed. 


We have already observed that Yaksas and Yaksis, though deities 
of vegetation, are constantly, though not invariably, provided with 
supports representing mythical aqueous animals, notably the inakara, 
more rarely the fish-tailed horse ( jala-furaga), elephant (jalebha), 
or lion, or the fiower of a lotus; and we have naturally assumed that 
this is an indication of the intimate connection of these deities of fer- 
tility with the life-giving Waters. Figures of this kind occur not only 
singly on pillars, pilasters, and stelae of various kinds, but also in 
pairs (affronted or addorsed) as bracket figures supporting the archi- 
traves of structural foranas; such pairs without vehicles are found on 
the SaiicT toranas, with elephant or other vehicles at the Kaiikali Tila, 
Alathura, but for our purpose the Bharhut example (chamfer reliefs, 
two sdlahhanjika Yaksis supported by lotus flowers),' and another 
from the Kankali Tila site (architrave bracket fragments, two female 
figures supported by inakaras)' are more significant. 

Later, towards the close of the Gupta period and thereafter, we 
meet with similar pairs of goddesses, sometimes with identical, some- 
times with differentiated vehicles, placed at the bases of the jambs of 
doorways. The manner in which such figures have found their way 
from their originally functional position as architrave brackets, to 
that of dvarapalas at ground level can be clearly traced. The case of 

^ Unpublished part of a pillar, above the medallion reproduced by Cunningham, 
Stupa of Bharhut, pi. XXXIII, 4. This seems to be the only early example of 
a salabhanjika Yaksi supported by a lotus, but the type recurs later, e. g., on 
the verandah pillars of the Ramesvaram cave, Elura (pi. 21, fig. 2, centre and 
right), and similarly at Badami, Cave IV. 

' Smith, Jaina stupa of Mathura, pi. XXXVI. 


the entrance doorway at Nasik, Cave III (ca. 130 A. D.) is especially 
instructive ; here it is very evident that the monolithic doorway effec- 
tively presents the projection of a structural tor ana against a flat 
wall surface ; only the space between the jambs, threshold, and lower 
architrave (now functioning as lintel), being perforated. It is true 
that in this instance the torana architrave brackets are rampant lions/ 
but we are nevertheless provided with the key to the origin of door- 
way forms such as those of the Candragupta cave, Udayagiri,^ where 
the architraval nature of the lintel is no longer recognizable, but there 
remain vestigial brackets consisting of paired sdlabhanjika Yaksis 
supported by makaras at the lintel level. Very numerous examples of 
the same kind, both goddesses standing on makaras, may be seen at 
Ajanta (Caves i, 5, 7, etc.) and at Bagh, and this seems to be the 
general rule in the Gupta period. The well-known Besnagar exam- 
ple, from a structural temple, and now in Boston {Yaksas, Pt. I, pi. 
14, 2) lacks a mate, but it may be assumed that it had once an exactly 
corresponding counterpart.^ 

Up to this point no change has taken place in the iconography, 
except that the makara has become more conspicuous, and that dwarf 
genii are often associated with it ; there can hardly as yet be any justi- 
fication for an identification of the twin figures as the individual god- 
desses Ganga and Yamuna. On the other hand at Deogarh * the god- 
desses are differentiated, one to the right being supported by a tor- 
toise, one to the left by a makara; at the same time the tree is now 
altogether omitted, there are umbrellas behind the heads of the god- 
desses, and each seems to hold or to have held a lotus bud. At Tigawa ^ 
the goddesses are similarly dift'erentiated by their vehicles, but the 
trees and salabhanjtkd pose are retained ; in other words, the type is 
transitional, combining the older and the new iconography. 

After this time the paired goddesses are transferred from their no 
longer functional position at lintel level, to a new position at ground 
level, where they function as dvdrapdlas, and henceforth throughout 

^ Corresponding lion brackets survive at Ajanta, Cave IV (A, S. W. I., IV, 
pi XXIV) ; and at the Amrta Cave, Udayagiri, where we have in addition 
salabhanjika Yaksis at lintel level, and river goddesses (the vehicles not recog- 
nizable) on the jambs. 

^ Vogel, J. Ph., Ganga et Yamuna dans Viconographie hrahmanique , £tudes 
Asiatiques, pi. 55. 

" However, no argument can be based on the fact that this is a figure from 
the right side; for at Bhumara and Aihole the usual relations of the differen- 
tiated goddesses are reversed. 

^Burgess, Ancient Monuments, pi. 249: here pi. 21, fig. i. 

^ Cunningham, A. S., Reports, IX, p. 46. 


the medieval period and subsequently, they appear in this position. . 
But while in northern India the jamb figures are generally differen- 
tiated as at Deogarh, in southern India, e. g., at Tadpatri ( Yaksas, 
Pt. I, pi. 19, i) the scheme of the twin figures, both supported by 
iimkaras, is retained. Longhurst,^ pointing out that the undift'eren- 
tiated twin figures are not met with south of the Ganjam District, 
calls them " duplicate figures of Ganga " ; but inasmuch as the tree 
and salabhanjikd scheme is always preserved, that even the dohada 
motif may be retained as at the Subrahmanya temple, Tan j ore 
{Yaksas, Pt. I, pi. 19, 2) (the motif also occurs on a pilaster of the 
Raj Rani, Bhuvanesvara), and that the goddess is not provided with 
a vase or lotus attributes or with an umbrella, it is, as pointed out by 
Vogel {loc. cit., pp. 397, 398), most questionable whether the desig- 
nation of river goddess is in any way appropriate. 

In the case of the differentiated northern types the correct desig- 
nation as individual river goddesses is placed beyond doubt (i) by 
the inscription on the Vaidyanatha temple, Baijnath, Kaiigra, in which 
the figures, still extant, are referred to as Gafiga and Yamuna,^ (2) by 
inscriptions at Bhera Ghat,^ and (3) by the description of Varuna 
in the Visnudharmottara, III, 52, where the river-goddesses attendant 
on him are called Gahga and Yamuna, and are said to be supported 
by a makara and a tortoise respectively. We have seen reason to think 
that these animals were originally symbols or forms of Varuna him- 
self, and it should be borne in mind that the rivers are at all times 
spoken of as his consorts. 

This leads to a consideration of the special case of the magnificent 
compositions found at each end of the Varaha shrine at Udayagiri. 
Each of these represents the flowing of two rivers into the sea, in 
which stands Varuna himself, holding his ratna-pdtra. These rivers 
are evidently the Ganges and Jamna, for goddesses are represented 
standing in each, holding vases, and supported respectively by a 
nmkara and a tortoise. Between the rivers there is dancing and music 
(nrta-gita-vada) . The sculpture is generally dated about 400 A. D. 
and thus represents probably the earliest known representation of 
the differentiated goddesses. I see no reason for post-dating the 
Varaha sculpture merely because the goddesses are differentiated; 
the less so, inasmuch as the headdress worn by Varuna is not far 
removed from Kusana types. It is true that the change in iconography 

^ Longhurst, Hanipi ruins, p. 116. 

■ Vogel, Ganga et Yamuna . . . . , pp. 387, 388. 

" Cunningham, A. S., Reports, IX, pp. 66-69. 


on shrine doorways seems to be taking place rather towards the end 
than at the beginning of the fifth century ; but the time interval is not 
excessive, and it may even be the case that the doorway types were 
affected and changed under the influence of just such representations 
as those of the Varaha relief, or of its literary sources. 

The mediaeval examples of the differentiated type on door jambs 
are very numerous/ The iconography is typically illustrated at 
Kharod, where each of the goddesses carries a vase of plenty (pilrna- 
kalasa) at shoulder level, and is provided with a dwarf attendant and 
an umbrella; and at Bajaura, where each carries a vase of plenty and 
a long-stemmed full-blown lotus flower, and is provided with a dwarf 
umbrella-bearer. But at Bajaura the goddesses both stand on ex- 
panded lotus pedestals, which are supported by their vehicles, makara 
and tortoise, though these are almost dissolved in decorative scroll 
work, and can hardly be distinguished. 

Some other representations of river goddess may be more briefly 
mentioned. At Elura there is a well-known shrine of three river god- 
desses, evidently Gahga, Yamuna, and Sarasvati; the supports or 
vehicles are not well preserved, but seem to be a makara, a tortoise, 
and an expanded lotus.^ At the Caunsat Jogini temple, Bhera Ghat, 
there are inscribed figures of Jahnavi (Ganga) with a makara, 
Yamuna, with a tortoise, and Uha, perhaps the Sarasvati, with a pea- 
cock.' A figure certainly representing Gafiga occurs on one of the pil- 
lars from Candimau, Bihar; the goddess rides on a makara with a 
floriated tail, and bows towards Siva, seated on a mountain before her ; 
behind her is an attendant holding the usual long-handled umbrella ; 
above the attendant, in the air, is an unrecognizable object resembling 
a bull's head (pi. 48, fig. i ) .* 

The river goddess Narbada or Rksini occurs with a makara pedes- 
tal at Tewar and Bhera Ghat, both on the river of the same name." 

^For Kharod and Bajaura, see Vogel, Ganga et Yamuna, pis. 52-54. Other 
noteworthy examples include those at Bhumara (Gupta), Mem. A. S. I., 16, 
1924, pi. Ill, a; the Lad Khan and some other temples at Aihole (A. S. L, 
A. R., 1907-08, pp. 191, 202, and pis. LXXXI, LXXXIX) ; Ramesvaram, Elura 
(A. S. I., N. I. S., V, p. 39 and pi. V) ; unknown source, Diez, Zzvei unhekannte 
Werke der indischen Plastik . . . . , Wiener Beitrage, I, 1926. 

' Burgess, in A. S. I., N. I. S., V, p. 34 and fig. 16. 

^ Cunningham, A. S. Reports, IX, pp. 66-69. 

*Banerji, R. D., Four sculptures from Candimau, A. S. L, A. R,, 1911-12, 
pi. LXXIV, I. Banerji's pillars (2) and (3) are parts of one and the same 
pillar ; his fig. i represents the right hand side of fig. 3 on the same plate. The 
pillars show also, in the lunettes, excellent examples of kinnaras and klrttimukhas. 

^ Cunningham, ib., p. 67. 


A Gupta coin already alluded to bears the figure of a goddess on a 
makara; she may or may not be Ganga. In eighteenth century Rajput 
paintings Gahga is represented as a four-armed goddess seated on a 
fish. On the other hand, when represented in Siva's matted locks (as 
usually in Nataraja images), Gahga is represented in the form of a 
mermaid (cf . gahgavatarana, Elura, in A. S. I., N. I. S., V, pi. XXVI, 
i). Some female figures with a makara vehicle certainly represent 
Parvati {e. g., at Elura, ib., pi. XXI, 2). 

So far we have considered no evidence for any kind of representa- 
tion of river goddesses previous to the Gupta period. However, there 
can be cited from Amaravati no less than four reliefs in which river 
goddesses {nadl-dcvata) are represented; in all cases they accompany 
or attend upon a Naga. Three "^ are found in representations of 
Kalika's homage to the Bodhisattva as he emerges from his bath in 
the Neranjana ; in each case there is a group of river-nymphs bear- 
ing vases (either plain, or of the purna kalasa type) on their heads 
or in their hands, w^ith which to do honor to the Bodhisattva. In the 
fourth instance ^ there are paired pseudo-chamfer reliefs on a railing 
pillar, representing two similar nymphs, each standing on a fish (pos- 
sibly intended for a makara) and bearing a tray of food at shoulder 
level, and a water vessel carried horizontally ; the two are approaching 
a theriomorphic Naga, who, coiled amongst lotuses, occupies the cen- 
tral panel of the triptych. The type is of interest from several points 
of view: (i) it preserves the old formula of paired representations 
on chamfers, (2) it gives us an undoubted example of duplicate river- 
goddesses supported by fish or makara, (3) the type bringing food 
and water connects with that of certain undoubted Yaksis (pi. 45, 
fig. 2), with the Deokali caryatide (pi. 18, fig. 3), and less directly 
with the Sri-Laksml type of Saiici, stupa 11,^ and (4) the water-vessel 
carried horizontally here and in several of the types just cited, con- 
nects with certain representations of the female genii of springs to be 
referred to below. Thus, the types of the undifferentiated river-god- 
desses connect on the one hand with those of Yaksis or dryads ; on 
the other hand, they exhibit in the vase attribute what may well have 
been the immediate source of this motif as it appears held by the 

^ PI. 26, fig. I : Museum of Fine Arts Bulletin, No. 160, p. 21 ; Vogel, Indian 
serpent lore, pi. VII, a, on the left side. 

- PI. 19, fig. I. 

^ Early Indian iconography, Pt. II, Srl-Laksmi, Eastern Art, I, 1929, fig. 16; 
and here pi. 14, fig. i, center. But here the food and drink are carried by- 


differentiated river-goddesses at the close of the Gupta period and 

One other type remains to be considered, that of the presiding 
female divinity of a sacred spring. Reference may first be made to 
the goddess of a sacred source at Majapahit, Java (pi. 23, fig. 3) ; ^ she 
holds in her right hand a bunch of lotuses, and under her left arm a 
v^rater vessel in the horizontal position above alluded to, and it cannot 
be doubted that this vessel represents the ever-flowing waters of the 
source over which she presides. To find a close parallel in India for 
this type we have to go back to Gandharan art of perhaps the second 
or third century A. D. : it is a rather mysterious fact that the Javanese 
goddess of a spring exhibits a resemblance, amounting to identity, 
with that of the flower girl Bhadra or Prakrti in several renderings 
of the Dipahkara legend; she, too, carries (pi. 23, fig. i) ^ a bunch of 
lotuses in her right hand, and a water vessel held horizontally under 
her left arm, precisely as in the Majapahit relief. The flower girl is 
apparently a human being; she becomes Prince Megha's wife, and 
remains his wife in all future incarnations up to the attainment of 
Buddhahood. Still it is difficult to ignore the significance of the names 
of the chief characters ; a girl named Abundance, or Nature, marries 
a prince named Cloud! Is it not possible that some older story or 
myth has here been adapted to Buddhist ends ? One observes also that 
the flower girl is found in a marshy place, as might be expected of 
the divinity of a pond or spring, though it is not necessary to the story ; 
and that she makes her appearance very opportunely, as divinities are 
apt to do when edifying purposes are to be accomplished. Or in any 
case the prototype adopted by the sculptor may have been that of a 
water nymph. 

One quite different type of the divinity of a spring has been found 
at Jagatsukh, Kulu (pi. 23, fig. 2) ; here the goddess holds a large 
vase perforated horizontally from back to front to permit the issue 
of water, and she stands on a makara. 

^Krom, N. J., L'Art javanais . . . . , Ars Asiatica, VIII, 926, p. 61, and 
pi. XXXII. Referring to the " tomb tanks " of Erlangga and of Udayana on 
Mt. Penanggungan, Stutterheim, in J. A. O. S., vol. 51 (in press), remarks 
" sometimes the water spouts from the breasts of a goddess, sometimes from 
the amrta-jar and often the whole scene is decorated with representations from 
the story of Garuda and the amrta." 

^ For the whole composition, see HIIA., fig. 92. 

Other illustrations of the legend, A. S. I., A. R., 1907-08, pi. XLII, d, and 
1909-10, pi. XVI, c; Burgess, Ancient monuments, pi. 140; cf. Foucher, L'Art 
greco-houddhique du Gandhdra, I, p. 27^ ff. In some cases the water vessel is 
carried upright. The purely Indian examples are fragmentary or doubtful 
(Vogel, in A. S. I., A. R., 1909-10, fig. 5 and pi. XXV, c). 



Plate i 

The well-known kalpa-vrksa capital of a dhvaja-stambha from Besnagar, usually 
dated in the third century B. C. The wishing-tree is a banyan (nyagro- 
dha), and between the hanging aerial roots will be seen a pot, and two 
bags, overflowing with money: on the other side of the tree, in a cor- 
responding position, are found a lotus flower and a conch each similarly 
exuding coins. These last (shown in detail on pi. 47, figs, i, 2) are clearly 
the nidhis (sankha and padma) of Kubera. Height 5 ft. 8 in.; Calcutta 
Museum. India Office photograph. 

Plate 2 

1. Seated pot-bellied Yaksa, with curly hair and moustache; he wears a dhoti, 

and is seated with a supporting patta encircling stomach and left knee. 
Height 3 ft. 8 in. From near Pali Khera, Mathura, now C3 in the 
Mathura Museum. Archaeological Survey photograph. 

2. Fragment of a railing pillar. Yaksa under a mango-tree ; perhaps Kamadeva. 

Height 2 ft. 7i in. 2nd century A. D. From the Chaubara mounds, 
Mathura, now J7 in the Alathura Museum. A. K, C. photograph. 

3. Fragment of a railing pillar (from a stairway). Yaksi under tree, with 

vessels of food and drink. Perhaps a form of Srl-Laksml, cf. my Early 
Indian iconography, 2. Srl-Laksml, in Eastern Art, i, 1928, figs. 16, 28, 
and B, C. 2nd Century A. D. Mathura Museum. A. K. C. photograph. 

Plate 3 

Relief from Nagarjunikonda, detail. Standing Buddha with the Yaksa Vajrapani 
below his right arm ; Yaksa dvdrapala with cdmara on lower left ; on the 
left, above, as coping relief, a Yaksa supporting a lotus-rhizome garland 
drawn from the open jaws of a makara. Third century A. D. Metropolitan 
Museum of Art, New York. Photograph by the same. 

Plate 4 

1. Mithnna (Yaksa and Yaksi) under a tree, supported by addorsed makara 

heads ; from the j-aws of one of the latter there emerges a lion. For the 
paired addorsed makara heads, cf. pi. 37, fig. 5, and pi. 39, fig. 2. On the 
lower left will be noticed a garuda bracket. From the right end of an 
architrave from Nagarjunikonda, third century A. D. Madras Museum. 

2. Kubera. Cup in right hand, purse in left (cf. pi. 8, fig. i, right hand figure), 

Kusana; Mathura. Victoria and Albert Museum, London. 

^ When "Amaravati " is given as source, it is to be understood that the draw- 
ings have been made from the various well-known publications of Amaravati 
reliefs, or more often directly from the series of photos taken by M. Goloubew 
in the Madras Museum. Nearly all of the drawings have been made by Mr. 
Sunchiro Tomita. 


Plate 5 

1. Relief: Hariti and Pancika, with some of their 500 children. Ajanta, Cave 11. 

About SCO A. D. 

2. Makara, with lotus rhizome. From a Sinhalese knife, eighteenth century. 

Plate 6 

medieval jaina images of kubera and his consort 

I, 2. Kubera and Bhadra, relief images in the Camundaraya Basti, Sravana 
Belgola. His attributes seem to be a lotus and a citron, hers perhaps the 
same. The trees, though considerably conventionalized, seem to be banyan 
and mango. Eleventh century. Mysore Arch. Surv. photograph. 

3. Kubera and Bhadra seated on a bench under a tree, each with a child in the 

arm. Money pots below the bench. Small seated image of Parsvanatha 
above the fork of the tree. About tenth to eleventh century. From 
Maldeh. India Museum photograph. 

Plate 7 

1. Sdtabhanjikd figure, viz. Yaksl under a mango tree, forming an architrave 

bracket of the north torana, Saiici. 

2. Lotus altar for hali offerings to Yaksas (cf. p. 5). From Anuradhapura, 

Early Medieval, Colombo Museum. 

Plate 8 

1. Trinity of Fortune: Ganesa, LaksmI (abhiseka) and Kubera, seated on lotus 

seats on a common stem. About eighth century A. D. University Museum, 
Philadelphia. Photograph by the same. See The University Museum 
Bulletin, Vol. 2, 1930, p. 15. 

2. Yaksa Atlantes; verandah of Cave III, Nasik. Early second century A. D. 

Cf. Pethavatthu Atthakatha, 45, 55, where Yaksas are called Bhumma 
deva, " Earth gods." Cf. also Yaksas, I, pi. 13, and Cunningham, Bharhut, 
pi. XV. 

Plate 9 

I, 2, 3. Pilasters from the pakara slabs, Jaggayyapeta. Left, Yaksi on fish- 
tailed elephant (jalebha or jala-hastin) = pi. 43, fig. 4 ; center, Yaksi on 
fish- tailed horse (jaia-turaga) ; right, padmapani Yaksa on fish-tailed lion. 
Second century B. C. Madras Museum. India office photographs. 

Plate 10 

I. Carnda Yakhi, under tree, supported by a fish-tailed horse (jala-turaga) . 

Bharhut railing pillar, 150-175 B. C. Indian Museum, Calcutta. India 

Office photograph. 
2 Yaksi under a tree, supported on a cushion on a makara; from a railing pillar, 

Mathura District. First or second century A. D. Museum of Fine Arts, 

Boston. Photograph by the same. 


Plate ii 

I, 2, 3. Lotus rhizomes bearing flowers, fruits, garment, jewels, etc., and pro- 
ceeding from the mouth of a (sky-) elephant. The text cited above, p. 4, 
suggests that the spray is here conceived as " torn by Airavata from the 
Wishing-tree of Paradise." Bharhut coping, 150-175 B. C. Indian 
Museum, Calcutta. Photograph by the same. See p. 4. 

4. Visnu Anantasayana, and the Birth of Brahma. Cambodian, classical period. 
Collection of C. T. Loo, Paris. 

Plate 12 

1. Architrave, south tor ana, Sanci : Yaksas spouting lotus sprays (rhizome 

and flowers). Two Yaksas in the center holding jewelled garlands. 
India office photograph. 

2. Detail of fig. I. Goloubew photograph. 

3. Lotus rhizome borne by Yaksas, and decoratively treated, providing a transi- 

tional type between those of Bharhut and Safici, and AmaravatT, Ni in 
the Mathura Museum, Vogel, Catalogue . . . . , pi. IV. About 100 
A. D. A. S. I. photograph. 

4. The lotus rhizome now fully decorated, drawn from the jaws of a makara 

by a lotus-seated goddess, probably Sri-Laksml, and supported by a dwarf 
Yaksa. Note incidentally the cable moulding above the lotus petal course ; 
like the torus of a " bell-capital," this cable-moulding represents the 
stamens of the open lotus flower. Madras Museum. India Office photo- 

Plate 13 

1. Similar lotus rhizome, rising from the jaws of a makara. Sanci, east torana. 

Goloubew photograph. Cf. Smith, Jaina stupa of Mathura, pi. XXVI. 

2. Lotus rhizome, flower and jewel bearing, and framing human figures, two 

of which are seated on lotus flowers, also animals and birds. The whole 
spray originates from the navel of a seated dwarf Yaksa at the base of 
the pillar, no longer extant, but shown in Fergusson, Tree and serpent 
zn'orship, pi. VIII, and our reproduction, pi. 36, fig. i. SaficT, south torana. 
Goloubew photograph. 

Plate 14 

1. Center medallion: SrT-Laksm! or a nadl-devata, amongst lotuses, cf. pi. 2, 

fig. 3, and pi. 19, fig. I. Another figure in the doorway above. Right and 
left, below, lotus rhizome rising from the jaws of makaras, and above, 
from the navels of dwarf Yaksas. Jamb pillar of railing, stupa II, Safici, 
first century B. C. or A. D. A. K. C. photograph. 

2. Lotus rhizome rising from a full vase (purna-ghata) supported by a dwarf 

Yaksa. Stele at Vihare II, Polonnaruva, Ceylon, medieval. Cf. pi. 42, 
fig. I. A. S. C. photograph. 


Plate 15 

1. Yaksas, apparently returning the end of the ornamented lotus rhizome to the 

mouth of a dwarf Yaksa. Coping, Amaravati, ca. 200 A. D. Madras 
Museum. India Office photograph. 

2. Makara, with the ornamented lotus rhizome proceeding from its mouth, and 

dominated by a dwarf Yaksa. Coping, Amaravati, ca. 200 A. D. Madras 
Museum. India Office photograph. 

Plate 16 

1. Makara, limestone, length 21 in., property of K. Minassian, New York. 

There is a socket in the center of the belly, and an opening extended from 
the mouth to the socket. This has evidently been a makara standard or 
small makara dhvaja stamhha, and may have been connected with a 
temple of Pradyumna (= Kamadeva) or even Varuna. Mr. Minassian 
possesses another incomplete example of the same kind. Nothing is known 
of the source, except that both were obtained in northern India. A date 
about the third or second century B. C. may be conjectured. Museum of 
Fine Arts photograph. 

2. The makara of the makara-dhvaja stamhha from Besnagar^ (see pi. 45, 

fig. 3). Length 3 ft. Second century B. C. Gwalior Museum. Gwalior, 
A. S., photograph. 

3. Makara, in a railing cross-bar medallion, Mathura District, ca. second cen- 

tury B. C. Similar examples from Bharhut and Bodhgaya are known. 
Museum of Fine Arts, Boston. Photograph by the same. 

Plate 17 
Full vases : 

1. Full vase (purna-ghata) in center ; on each side a conch with a lotus spring- 

ing from it, a vardhamdnaka (powder box), and a lotus apparently with 
a flame rising from it. Early ninth century. Candi Sewu, Java. A. K. C. 
photograph. Similar conches with lotuses are used also at Borobudur in 
the scene representing the honoring of the Bodhi-trees previous to the 
Great Enlightenment. 

2. Richly developed full vase as part of a pillar, with dwarf Yaksas at the sides 

blowing conches. Eighth century. Indra Sabha, Elura. India Office 

3. A similar richly developed full vase as a pillar capital; verandah of Cave 

XXIV, Ajanta. Sixth century. India Office photograph. 

4. Wooden tracery window, with full vase with richly developed foliage. South- 

ern India, eighteenth century present position (probably still Chipping 
Campden, England) unknown. A. K. C. photograph. 

^ See Cunningham, A, S., Reports, X, pp. 42, 43, and pi. XIV ; Bhandarkar, 
D. R., in A. S. I., A. R., 1913-14, pp. 189-191 (that the makara had a rider is 
unlikely), and Archaeology and Vaisnava tradition, Mem. A. S. I., 5. 


Plate i8 

1. One side of the " Pali Khera group," C2 in the Mathura Museum. A Bac- 

chanalian Yaksa, probably Kubera,' seated on a mountain, attended by 
female cup-bearers, with trees behind, the whole forming the base sup- 
porting a bowl, of which only part is preserved. First or second century 
A. D. Indian Museum photograph. 

2. A similar pedestal which served as the support of a bowl, of which only a 

part is preserved. The two YaksTs stand under an asoka tree, which is 
fully represented at the back of the stone. One of the YaksTs holds a 
parrot. From Kota, near Mathura. Probably second century A. D. See 
Cunningham, A. S. Reports, XX, p. 50, and pi. III. India Museum 

3. Back and front of a female figure and column supporting a bowl. The female 

figure suggests SrI-Laksm! and (or) a Yaks!, cf. pi. 2, fig. 3 (=pl. 45, 
fig. 2) ; pi. 14, fig. I, center. 
Second century A. D. Bharata Kala Parisad, Benares. Photograph by the 

Plate 19 

1. Naga, amongst lotuses, i. e,, in water, and attended by nadi-devatds supported 

by fish or makaras. Detail from a railing pillar, AmaravatT, about 200 
A. D. Madras Museum. India Office photograph. 

2. Left hand part of a scene representing the Buddha's Bath in the Neranjana 

(for the whole see Vogel, Indian serpent-lore, pi. VII, a), showing nadl- 
dcvatas bringing offerings of full vases {punna-ghata) to the Bodhisattva. 
Detail of a railing pillar, Amaravati, about 200 A. D. British Museum. 
A. K. C. photograph. 

3. The Jamna River goddess, Yamuna Devi, supported on a tortoise, and at- 

tended by a dwarf with an umbrella, and maid with a basket. Both at- 
tendants are supported by fish. Gupta. From Paharpur, Rajshahi Dis- 
trict. A. S. I. photograph. 

Plate 20 
Reliefs at the two ends of the Varaha cave, Udayagiri, Gwalior. On the left, 
the two rivers Jamna and Ganges, with the river goddesses standing 
respectively on a tortoise and a makara, flowing into the ocean, wherein 
stands a deity, probably Varuna, with a vessel. Above the two god- 
desses, and between the rivers, a dancing scene, with the dancer ki the 
center, surrounded by players on the harp, lute, flute, and drums. Above 
this, an angel(?) with a garland (?). On the right, a similar composition, 
omitting the dancing scene. Ca. 400 A. D. India Office photograph 

Plate 21 

I. Details of the doorway of the Gupta Dasavatara temple, Deogarh. On the 
left the river goddess Gariga Devi, supported by a makara, an umbrella 
above her head ; on the right Yamuna Devi, supported by a tortoise, and 
with an umbrella over her head. About 600 A. D. India Office photograph. 

2 Left end of the verandah of the Ramesvara shrine, Elura. On the left, Ganga 
Devi, with a dwarf, and supported by a makara; center and right, pillars 
with piirna-ghata capitals, and saldbhanjikd brackets (Yaksis under 
trees). Seventh century. India Office photograph. 


Plate 22 

Yamuna Devi, supported by a tortoise, and standing amongst lotuses, under a 
makara-torana, the makaras with dwarf Yaksa riders. Kailasa, Elura; 
eighth century. Goloubew photograph. 

Plate 23 

1, Detail from a relief (see H. I. I. A., fig. 92) of the DTparikara Jataka. 

Gandhara, second century A. D. Prince Megha ("Qoud") with a purse 
is purchasing lotuses from Prakrti ("Nature"), alias Bhadra ("Plenty," 
also a name of the consort of Kubera) ; Prakrti holds the lotuses in her 
right hand, a vessel under her left arm (cf. fig. 3 on same plate). Prop- 
erty of K. Minassian, New York. 

2, Goddess of a spring, supported by a niakara; four armed, holding a large 

vessel perforated from front to back of the slab ; a camara and lotus held 
in the other hands. Height 28 in. In a temple dated 1428 A. D. at 
Jagatsukh, Kulu, but the sculpture is probably older. See A. S. I., A. R., 
1907-08, p. 267 and fig. 2; and Bhattacharya, B. C, Indian images, p. 44 
and pi. XXX, fig. i. Lahore Museum. A. S. I. photograph. 

3, 4. Two goddesses of a sacred spring at Mojokerto, Java, now in the Museum 

at Batavia ; both were " adossees au mur afin de verser a I'exterieur Teau 
des urnes qu'elles tiennent a la main" (Krom, N. J., in Ars Asiatica VIII, 
p. 61 and pi. XXXII). The resemblance between one of these (fig. 3) 
and the Prakrti of fig. i will be remarked. Height of fig. 3 is .72 m. 
After Krom, loc. cit. 

Plate 24 

Abhiseka of Sri-Laksmi, Ravana ka Khai, Elura; eighth century. There are 
four dig-gajas, or sky elephants. Right and left of the goddess are four- 
armed crowned male deities holding vessels; one of these, probably the 
one on the right who holds a conch in the upper left hand, must be Varuna. 
The predella composition represents a lotus pond, with Nagas amongst 
the lotuses, holding full vessels (purna-ghafa) . 

Plate 25 
Half-seen Yaksas in trees : 

1. A rukkha devatd offering food and drink; probably an illustration to the 

Story of the Treasurer, etc., Dhammapada Atthakatha, I, 204, see 
Burlingame, Buddhist legends, I, 277. From the Bharhut coping, 150-175 
B. C. Indian Museum, Calcutta. Cf. J. R. A. S., 1928, p. 393. 

2. Detail from a Parinirvana, with a rukkha-devata seen patresv ardhakaydn 

abhinirmaya (see Yaksas, I, p. 33, note i), or upaddha-sarira (Jataka VI, 
370). H8 in the Mathura Museum, Vogel, Catalogue, p. 129. 

3. As fig. I on this plate; from a railing pillar, Bodhgaya, about 100 B. C. 

4. Detail from the Sutasoma Jataka ; only the face of the rukkha devata is seen. 

From wall paintings at Degaldoruwa, nr. Kandy, Ceylon. Eighteenth 
century. See p. 7. 


Plate 26 

1. The Bodhisattva, after the Bath in the Neranjana, welcomed by the Naga 

Kalika, two Naginis and a Deva ; and above, left, four nadl-devatds with 
full vases, and other deities on the right. The Deva, apparently with 
matted locks, may be Brahma. The Buddha is represented by foot-marks 
on a lotus pedestal, and a fiery pillar surmounted by the ratnatraya (cf. 
SaiicI, north torana, left pillar, outer face). AmaravatT, about 200 A. D. 
After Fergusson, Tree and serpent worship, pi. LXVII. 

For analogous but more detailed representations of the same scene, see 
Vogel, Indian serpent lore, pi. VII, a (= in part our pi. 19, fig. i), 
Museum of Fine Arts Bulletin, No. 160, and Krom, Life of Buddha on 
the stupa of Barabudnr, fig. 91 (= in part our pi. 41, fig. 4). In all cases 
the nadi-devatds bear full vases, but in the Boston and Borobudur ex- 
amples without foliage. 

2. Presentation of the infant Bodhisattva before the Yaksa Sakyavardhana, who 

appears within a shrine. AmaravatT, about 200 A. D. After Fergusson, 
lac. cit. pi. LXIX. Cf. Yaksas, I, p. 42. 

Plate 27 

1. Offering vase with vegetation. Sumerian relief from Susa, third millennium 

B. C. Original in the Louvre; for the whole, see Unger, E., Sumerische 
nnd akkadische Kunst, 1926, fig. 59; and cf. Ward, Seal cylinders of 
Western Asia, No. 1235, and Heuzey, Origines orientales de Vart, 

2. Lower part of a kurnhha-pahjara, showing a full vase (purna-ghata) with 

a pilaster taking the place of the central vegetative motifs. Hazara 
Ramacandra temple, Hampi (Vijayanagar) ; begun A. D. 1513. 

Plate 28 

Full vases (purna-ghata or kalasa) : 

1. AmaravatT, about 200 A. D. 

2. Used as a welcome offering. Story of Sudhana and Ratnacuda, Borobudur, 

Java. About 800 A. D. Krom en Erp, Beschrijving . . . . , Series II, 
pi. XV, No. 30. Cf. our pi. 41, fig. 4. 

Plate 29 
Lotus : 

1. Rhizome with flowers, buds, and leaves, rising from a full vase. AmaravatT, 

about 200 A. D. or earlier. After Fergusson, Tree and serpent worship, 
pi. LXXXIX. 

2. The same motif, SaiicT, north torana, left pillar, outer face. 

3. The same motif, more formally treated, and combined with addorsed animals 

(as also commonly at SaficT). From a stele near the south vahalkada. 
eastern dagaba, Anuradhapura, Ceylon. 

4. Below, a lotus palmette; above, lotus rhizome springing from a makaras 

jaws, with hamsas perched on the flowers. From a railing pillar, Sarnath 
first century B. C. 

5. Rhizome with flowers, buds, and leaves, rising from the navel of a dwarf 

Yaksa. SaficT, first century. A detail from pi. 14, fig. i. 


Plate 30 
Lotus : 

1. Rhizome with foliage dissolved in arabesque, rising from the navel of a 

Yaksa. Detail from an early Pala door jamb, Indian Museum, Calcutta. 

2. Rhizome with vine and lotus elements, rising from the jaws of a makara. 

Safici, north torana, left pillar, outer face. 

3. Rhizome with flowers, buds, and leaves, rising from the mouth of a Yaksa. 

AmaravatT, about 200 A. D. or earlier. After Fergusson, Tree and ser- 
pent worship, pi. LXXXIX. 

Plate 31 
Full vases and other symbols : 

1. Detail from a Jaina tympanum, showing three vaddhamanakas , a panna- 

pacchi or panna-puta, and a punna-ghata. From the KahkalT TTla, 
Mathura now J 555 in the Lucknow Museum, see Smith, Jaina stilpa of 
Mathura, pi. XIX. 

2. Details from a Jaina ayagapata, showing the astamangaJa, from left to right, 

above, fish, mirror, sirivaccha, vaddhamanaka; and below, ratnatraya, 
panna-pacchi or -piita, bhadddsana (?), and punna-ghata. From the 
Kankali Tila, Mathura ; now J 249 in the Lucknow Museum ; see Smith, 
Jaina stupa of Mathura, pi. XC. 

3. Detail from a Jaina ayagapata, vine springing from a full vase (punna-ghata). 

From the Kankali Tlla, Mathura; now J 253 in the Lucknow Museum. 
See Smith, Jaina stiipa . . . . , pi. X. Cf. ib., pi, XXII. 

4. Lotus springing from a full vase, Amaravati. 

Plate 32 
Full vases: 

1. Full vase capital of a pilaster, Dasavatara Gupta temple, Deogarh. 

2. Full vase, from a railing pillar, Sarnath; first century B. C. Sahni, Cata- 

logue. No. D(a)i. 

3. Full vase, Amaravati: Fergusson, Tree and serpent worship, pi. LXXVII. 

4. Full vase, SaficT, first century B. C. 

5. Full vase, SaiicI, called early Maurya in A. S. I., A. R., 1906-07, p. 79 and 

pi. XXVIII. 

Plate 2)?) 

Full vases and Fountain of Life : 

1. Full vase, from an architrave, Mathura, second century A. D. M3 in the 

Mathura Museum; Vogel, Catalogue . . . . , p. 163, and A. S. I., A. R., 
1909-10, pi. XXVIII, a. 

2. Full vase, one of the astamangala, from a fifteenth century Jaina manuscript ; 

Hiitteman, Miniaturen zum Jinacarita, Baessler Archiv, 1914. 

3. Full vase, from an embroidered jacket, Sinhalese, nineteenth century. 

4. Full vase, held by a Naga dvarapala, Anuradhapura. 

5. Lions, emerging with vegetation from the jaws of makaras, and running 

towards a flowing full vase (Fountain of Life motif) ; AmaravatT, Ca. 
200 A. D. 


Plate 34 
Lotus and Yaksa : 

I, 2. Two examples of the lotus rhizome, with buds, flowers and leaves, rising 
from a dwarf Yaksa's mouth; Bharhut, 150-175 B. C Fig. 2 is restored. 

Plate 35 

Lotus and Yaksa: the juxtaposition of these examples from Safici and Ama- 
ravati alone suffice to show the identity of the earlier realistic lotus 
rhizome and the so-called garland of the Amaravati coping. For this 
identity cf. also pi. 40, figs, i, 2, 3. 

1. Safici, stiipa I, south torana, architrave; first century B. C. A lotus rhizome, 

bearing flowers, etc., springs from the dwarf Yaksa's grimacing mouth, 
narrow at first then swelling to the full size; the Yaksa's left hand rests 
on the first node with its scale leaf, while his right hand holds a pearl 
garland. Another lotus spray springs from his navel. For the whole 
composition, see pi. 12, fig. 2. Sir John Marshall calls these Yaksas 
k'lcakas, " spouting forth all summer." 

2. The same motif from the coping, Amaravati, about 200 A. D. As before the 

rhizome rapidly swells to its full thickness, but it is here elaborately 
decorated, which disguises its true character; the nodes, too, are decora- 
tively treated (and as shown on pi. 12, fig. 4, made symmetrical rather 
than realistic). The lower course shows a lotus rhizome with flowers 
rising from the jaws of a water elephant. The upper course shows a 
decorative treatment of the expanded lotus seen in profile, the cable motif, 
as already in Asokan capitals, representing the stamens (cf. Indian 
Historical Qtly., VI, p. 37^). 

Plate z^ 
Lotus rhizomes and Yaksas : 

1. Lotus rhizome rising from the navel of a dwarf Yaksa, from Fergusson, 

Tree and serp^ent worship, pi. VIII. This is the now missing base of the 
torana pillar shown on pi. 13, fig. 2. Cf. Smith, Jaina stiipa of Mathurd, 
pi. XXVI. 

2. Dwarf Yaksa with exaggerated penis, from a railing pillar, Bodhgaya; about 

100 B. C Cf. pi. 43, fig. 7. 

3. Dwarf Yaksa, with vine springing from the mouth, and held in the hand; 

from a stone bowl, Mathura, of Kusana date. See A. S. L, A. R., 1915-16, 
Pt. I, pi. V, d. 

4. Dwarf Yaksa with lotus rhizome apparently (since the direction of growth 

is from left to right) reentering his mouth. Amaravati, coping, first cen- 
tury B. C? 

5. Lotus rhizome with nodes and flowers, very simply treated; from the altar 

in the verandah of the old monastery at Bhaja, early second century B. C. 


Plate 2>7 

Makara and lotus; water-elephant (jalebha) and lotus. 

1. Klrttimukha (full-face makara), with vegetation springing from the jaws. 

From the Dasavatara Gupta temple at Deogarh. About 600 A. D. 

2. Lotus rhizome with flowers, leaves, and animals, drawn from the jaws of a 

makara by a dwarf Yaksa; another Yaksa is using an elephant-goad to 
open the makara' s jaws ; another makara is ridden by a Yaksa. AmaravatT, 
about 200 A. D. 

3. Makaras with interlocked tails; a lotus rhizome with flowers and leaves 

springs from the open jaws on the left, and following the direction of 
growth around the circumference of the medallion, reenters the jaws of 
the makara on the right. Amaravati, about 200 A. D. Cf. Smith, Jaina 
stupa of Mathurq, pi. X. 

4. Lotus rhizome with flowers, buds, and leaves, springing from the jaws of a 

water-elephant {jalebha). Amaravati, about 200 A. D, Compare pi. 40, 
fig. 4, and contrast pi. 11, figs, i and 3. 

5. Decorated lotus rhizome springing from makara jaws; a combination of the 

decorated node and source themes, treated symmetrically. Note also the 
rootlets, which hang from the node. Amaravati, coping; about 200 A. D. 
The motif seems to occur first at Mathura, see Smith, Jaina stttpa of 
Mathura, pi. IX. For the addorsed makara heads, cf. pi. 4, fig. i, and 

pi. 39, fig. 2. 

Plate 38 

I. 2, 3. Lotus rhizome, with leaves, flowers and buds, running between af- 
fronted makaras. From the direction of growth it will be seen that the 
spray rises from the jaws of one and enters the jaws of the other makara. 

Plate 39 

1. Lotus rhizome with flowers, buds, fruits and leaves, from the hammiyd rail- 

ing of Stupa IV, Sanci; second century B. C. Length 5 ft. 7 inches. 
A 69 in the Safici Museum. Catalogue, p. 28 and pi. XL 

2. Detail from a "moonstone," Polonnaruva: two pairs of affronted makara 

heads with lotus sprays, and between them a klrttimukha. About the 
twelfth century. 

3. Lotus spray with fully developed acanthiform leaves, springing from a 

makara. Amaravati, about 200 A. D. 

Plate 40 

Lotus ; lotus palmettes and jalebha. 

I, 2, 3. Nodes of the lotus rhizome at Bharhut, Safici, and Amaravati. The 
two first are treated realistically, preserving the direction of growth; in 
the third, while the pedicule motif is retained, the leaf is duplicated for 
the sake of symmetry, without regard to the direction of growth. 

4. Detail from a tympanum arch. Ram Gumpha, Udayagiri, first centurj- 

B. C. : lotus palmettes rising from lotus leaves, and flowers, forming 
a continuous spray springing from the jaws of a water-elephant 
(jalebha). Cambridge History of India, I, pi. XXVIII. 


Plate 41 

1. Merman with double fish tail and elephant ears ; the latter like those of some 

dwarf Yaksas, cf. pi. 43, fig. 7. Sarnath, railing pillar, first century B. C. : 
Sahni, Catalogue .... D(a)6, p. 209, and pi. VI. Cf. Smith, Jaina 
stupa of Mathura, pi. IX, similar figures but female. 

2. Merman with double fish tail, each fork terminating in winged dragons. 

Mathura, about 100 A. D. 

3. Design of ever-flowing vases, built up from the common Sumerian symbol of 

the vase and two rivers. After Cros, G., Nouveaux fouilles de Tello, 
1910, pi. VIII, fig. 2. 

4. The Bodhisattva welcomed by Brahma with a full vase, on emerging from 

the Neranjana, Borobudur, Java. For the whole, see Krom, The life of 
Buddha on the stupa of Barabudur, fig. 91. Cf. Jdtaka, I, 93, where mer- 
chant's daughters welcome the Buddha in the same way, also pi. 19, fig. 2. 

Plate 42 

1. Lotus rhizome with flowers and leaves framing animals, etc., rising from 

a full vase supported by a dwarf Yaksa. Amaravati, about 200 A. D. 

2. Drawing from pi. 9, fig. 2, pilaster with YaksT, supported by a water horse 

(jalaturaga) ; Jaggayyapeta, second century B. C. 

Plate 43 

1. Water bull, from a railing pillar, Bharhut, 150-175 B. C. Note the small fins. 

2. Water horse (jalaturaga), vehicle of Cafnda YakkhT, Bharhut (see pi. 10, 

fig. i). About 150-175 B. C. Note the small fins. 

3. Ratnatraya symbol, the two horns composed of makaras. Amaravati, about 

200 A. D. Cf. Smith, Jahm stupa of Mathura, pi. L, 2. 

4. Water elephant (jalebha), vehicle of a Yakkhi, see pi. 9, fig. 2, and pi. 42, 

fig. 2. The best example of a jalebha or jala-dvipa will be found in Smith, 
Jaina stupa of Mathura, pi. LXXIII, fig. i. 

5. Typical lotus palmette, composed of leaves, tendrils and flowers, rising from 

a leaf at a node, cf. pi. 40, fig. 4. Safici, east torana. 

6. Makara with open jaws, forming a soma siltra of a Siva temple; a warrior 

standing in (emerging from ?) the open jaws. Hoysala. Mysore A. S., 
1913-14, pi. V. 
7 Dwarf Yaksa, with elephant ears and exaggerated penis, dragging at the 
upper jaw of a makara. Bodhgaya coping, about 100 B. C. 

Plate 44 

1. Abhiscka of Sri-LaksmT. Tympanum of a doorway, Ananta Gumpha, Udaya- 

giri, Orissa. First century B. C. 

2. Yaksa with lotus sprays, seated on lotus; larger rhizome sprays with flowers, 

etc., and hamsas perched on leaves, extending to right and left. Detail 
from the Dhamekh stupa, Sarnath, sixth century A. D. 

3. Lotus pedestal of a standing figure, chamfer relief on a Bharhut railing 

pillar; showing the petals, stamens, and pericarp, corresponding to the 
petals, cable moulding, and abacus of a lotus (so-called "bell") capital. 

4. Similar pedestal, but of a larger figure (unpublished) ; Bharhut. The com- 

position is here of the lotus-palmette type, with a scale leaf, indicating a 
node, at the center of the lower margin, cf. pi. 43, fig. 5. 


Plate 45 

1. Varuna, seated with his makara, as Guardian of the soma, which seems to 

be established on a rock (cf. " Varuna has placed Agni in the waters, the 
Sun in heaven, Soma on the rock," RV., V, 85, 2). From the composi- 
tions illustrating the Rape of the Soma, Cave IV, Badami, sixth century ; 
after Chanda, Bas-reliefs of Badami, Mem. A. S. I., 25, 1928, pi, XXIII. 
Another representation of Varuna with the makara will be found ib., 
pi. XXI, c. 

2. As pi. 2, fig. 3. 

3. As pi. 16, fig. 2, but showing the makara in place; after Cunningham, loc. cit. 

Plate 46 

The con^ch (sahkha) : 

I, 2. The sankha and padma nidhis of Kubera ; details from pi. I, on the right: 
both are exuding coins, 

3. A winged sahkha, similarly exuding coins or pearls ; medallion of a railing 

cross bar, Mathura. After Smith, Jciina stnpa of MathurCi, pi. LXXI, 7. 
In view of the rarity of the winged sankha symbol, attention may be 
called to a winged sahkha standard at Borobudur (Krom en Erp, 
Beschrijving . . . . , Series II, pi, IX, No. 18), 

4. Sahkha with lotus; detail of pi. 17, fig. i. 

5. Lotus rhizome, dissolving into arabesque, springing from a sahkha. From 

Mathura, probably of Gupta date. After Smith, JCiina sffipa of Mathura, 

6. The same, with the nodes now hardly recognizable, from the Dasavatara 

Gupta temple at Deogarh. 

7. Plaque with sahkha, and lotuses in angles ; from Basarh, of Kusana or Gupta 

date. A. S. I., A. R., 1903-04, p. 98, fig. 10. 

Plate 47 

1. Birth of Brahma, from a lotus springing from the navel of Narayana (Visnu- 

Anantasayana). Dasavatara, Eltira, eighth century. See my The Tree 
of Jesse and Indian sources of parallels, Art Bulletin, Vol. XI, 1929, 
and above, p. 2. 
" Makara " motifs in European art : 

2. Klrttimukha type, from a loth century Psalter, British Museum, Harley 

MS. 2904; Millar, E. J., English illuminated manuscripts from the Xth to 
the Xlllth century, 1926, pi. XI. 

3. Makara and lotus rhizome type, from the doorway of Aal church in Halling- 

dal, Norway, now in Christiania (Oslo) University. About 1200 A. D. 
From the cast in the Victoria and Albert Museum, London. 


Plate 48 

1. Gariga Devi approaching Siva. Candimau, Bihar, fifth-sixth century A. D. 

[=A. S. I., A. R., 1911-12, pi. LXXIV, figs. I and 3]. See p. 69. 
A. S. I. photo. 

2. Makara with lotus: Bharhut, ca. 175 B. C. Indian Museum, Calcutta: photo- 

graph by the same. 

3. Abhiseka of Sri-Laksmi : beside her a dwarf Yaksa seated, supporting a bowl, 

and a pillar surmounted by a cock. From Lala Bhagat, Cawnpore District : 
Suiiga, second century B. C. Lucknow Museum : photo by the same. See 
Prayag Dayal, A note on Lala Bhagat pillar, Journ. U. P. Hist. Soc. 
IV, 2, 1930, p. 38. 

Plate 49 

Sri-Laksmi, on and amongst lotuses rising from a " full-vessel." From Mathura, 
now B89 in the Lucknow Museum. H. 3 ft. lo? inches. Kusana, ca. 
2"'^ century A. D. 

Plate 50 

Details from the coping, Bodhgaya railing ca. 100 B. C. Sea monsters 
(makaras) and dwarf Yaksas, etc. 






I "' ^^^^<> 


^^i^^ ■'"^, *i^^^^»#^, ' %■ ± . ^ 'j^'^... 

*& .s ««ii* ^-^ 

Relief from Nagarjunikonda, showing Vajrapani beside the Buddha; and two 

other Yaksas. 



I. HaritI and Pancika : Ajanta, Cave 11, 

2. Makara and lotus spray: Ceylon, i8th century. 






• >^f V 


.<'/v -.#2*^',;,i 

o ^ 

-^ o 
3 '^ 



I. Trinity of Furtuiie : danesa, Laksmi, Kubera. University Museum, 


2. Yaksa Atlantes, Nasik, Cave III. 






|/ _!^pm^. 

I. Camda Yakhi, Bharhut. 

2. Yaksi, Mathura. 


■ \/ li K' y >,' 


swt.^^.iiis-'*! ,.j, >/^.-»cm- 

I, 2, 3. Lotus rhizomes proceeding from the mouth of an elephant (Aira- 
vata) ; coping, Bharhut. 

4. Visnu recHning and Birth of Brahma, Cambodian. 



I. Architrave, south torana, Sanci. 

2. Detail of Fig. i, SancT. 

3. Lotus rhizome and Yaksas, Mathura. 

4. Detail, coping, Amaravati. 



1. Detail of pillar, East gate, Sanci. 

2. Detail of pillar. South gate, Safici : for the original base see pi, ;i6, 

fig. I. 



msm^^^- ^"^^WW^^ 


1. Jamb pillar of railing: Stupa II, Sanci. 

2. Stele, Polonnaruva. Ceylon. 




rt G 

c/;- 03 

^ - 

rt ^ 






rt M-t 


^ b/3 

^ E 


-^ o 

^ a 


C Q^ 


N P 


IS o 


S^ _N 

-3 1-1 

O '^ 

-^ 2 

-13 o 

CU ^ 

E '^ 

Qj a-i 

P "i=i 


n3 "a 

i:^^ i-- ,„#i!»-ii 

^ WJ' 




I. Panel, Candi Sewu, Java. 

2. Detail of pillar, Indra Sabha, Elura 

3. Detail of pillar. Cave XXIV, Ajanta 

4. Tracery window. Southern India. 



I. From Pali Kher; 

2. From Kota. 

3. From Mathura District. 
Three pedestals supporting bowls. 


I. Naga and )iadl-dcz'atds. Amaravati 

2. Nad'i-devatas, Amaravati. 

3. Yamuna Devi, Paharpur. 





The Jumna and Ganges entering the ocean: Udayagiri, Gwalior. 



I. Upper angles of doorway, Gupta temple, Deogarh. 

2. Left end of verandah, Ramesvara shrine, Elura. 



Yamuna Devi, Elura. 



I. Dipankara J., detail. 

2. Goddess, Kulu. 

3, 4. Two goddesses of a spring, Java. 


AbJitseka of Sri-Laksml, Ravana ka Khai, Elura. 



I. Story of the Treasurer, etc., 

2. Detail, Parinirvana 

3. Story of the Treasurer, etc. ; Bodhgaya. 4. Detail, Sutasoma Jataka, Ceylon. 
Half seen Yaksas in trees. 



a; 03 

^ > 

+-- 103 

o u 


n ^ 






1. Bihar. 

2. Saiici. 


3. Amaravati. 



I. Auspicious symbols, Mathuri 

2. Eight auspicious symbols, Mathura. 

3. Vine and vase, Mathura. 

4. Lotus and vase, Amaravati. 



I. Deo.oarh. 


•*/YVVVr\liiT\Jlli ivif).iii)Lir. 

3. Ainaravati. 

2. Sarnath. 

4. Sanci. 

Full vases ( punna-gliafa). 

5. Sanci. 



I. Mathura. 

From a Jaina MS. 

3. Sinhalese embroidery. 

4. Anuradhapura. 

5. Fountain of Life : Amaravati. 
Full vases (piinna-ghata). 



I. Pillar base. 

2. Railing medallion. 
Lotus and Yaksas, Bharhut. 



I. Sanci, architrave (see pi. 12, fig. i), 

2. Amaravati coping. 
Lotus and Yaksa. 



I. Safici. 

2. Bodligaya. 

3. Mathura. 

4. AmaravatT. 

5. Bhaja. 
Lotus rhizomes and Yaksas. 



4. Amaravati. 

5. Amaravati. 
Lotus and inakara, lotus and jalcbha. 















I. AmaravatT. 

2. Jaggayyapeta. 


PLAT b 43 

I. Bharhut. 

2. Bharhut. 

4. Jaggayyapeta. 

3. Amaravati. 

5, Sanci. 

6. Mysore. 

7. Bodhgaya. 



I. Abhiscka of Sri, Udayagiri. 

i. From Dhamekh stupa, Sarnath. 

3 and 4. Bharhut. 



I. Varuna, Badaml. 

2. Mathura (= pi. 2, fig. 3). 3. Besnagar (=pl. 16, fig. 2). 



1 2 

I and 2. Besnao-ar. 

3- Mathura. 



5. Mathura 

4. Java. 

6. Deogarh. 7. Basarh. 

The conch (saiikha) ; and lotus. 


I. Birth of Brahma, Elura. 

2. European, loth century. 3- European, ca. 1200 A. D. 


I. Gaiiga Devi approaching Siva; Candimau. 

3. Sri-Laksmi ; Lala Bhagat. 




S'ri-LaksmT, Mathura : Lucknow Museum. 






3 ^Dflfl DDMBflflflT fl 

nhanth BL1215.Y3C77 
V. 2 Yaksas ...